Salilpro

Measurement and scaling: fundamentals, comparative and non-comparative scaling

2010-06-06T09:07:00.000-07:00

Stage 1 Problem definition
Stage 2 Research approach developed
Stage 3 Research design developed
Stage 4 Fieldwork or data collection
Stage 5 Data preparation and analysis
Stage 6 Report preparation and presentation
Objectives
After reading this chapter, you should be able to:
1. introduce the concepts of measurement and scaling and show how scaling may be considered an extension of measurement;
2. discuss the primary scales of measurement and differentiate nominal, ordinal, interval and ratio scales;
3. classify and discuss scaling techniques as comparative and non-comparative and describe the comparative techniques of paired comparison, rank order, constant sum and Q-sort scaling;
4. explain the concept of verbal protocols and discuss how they could be employed to measure consumer response to advertising;
5. describe the non-comparative scaling techniques, distinguish between continuous and itemised rating scales, and explain Likert, semantic differential and Stapel scales;
6. discuss the decisions involved in constructing itemised rating scales;
7. discuss the criteria used for scale evaluation and explain how to assess reliability, validity and generalisability;
8. discuss the considerations involved in implementing the primary scales of measurement in an international setting;
9. understand the ethical issues involved in selecting scales of measurement.
‘When you can measure what you are speaking about and express it in numbers, you know something about it’. – Lord Kelvin
Overview
Once the marketing researcher has a clear understanding of what they wish to understand in their target respondents, they should consider the concepts of scaling and measurement. These concepts are vital in developing questionnaires or ‘instruments of measurement’ that will fulfil their research objectives in the most accurate manner. This chapter describes the concepts of scaling and measurement and discusses four primary scales of measurement: nominal, ordinal, interval and ratio. We describe and illustrate both comparative and non-comparative scaling techniques in detail. The comparative techniques, consisting of paired comparison, rank order, constant sum and Q-sort scaling, are discussed and illustrated with examples. The non-comparative techniques are composed of continuous and itemised rating scales. We discuss and illustrate the popular itemised rating scales – the Likert, semantic differential and Stapel scales – as well as the construction of multi-item rating scales. We show how scaling techniques should be evaluated in terms of reliability and validity and consider how the researcher selects a particular scaling technique. Mathematically derived scales are also presented. The considerations involved in implementing scaling techniques when researching international markets are discussed. The chapter concludes with a discussion of several ethical issues that arise in scale construction. We begin with an example of how the use of different types of scale can give quite different powers of analysis and interpretation.
[Photo near hear]
example
Numbers, rankings and ratings: France is on top
According to the international football federation (FIFA) (www.fifa.com) 2005 end-of-year rankings, world champions Brazil maintained their supremacy at the top with 838 points and the Czech Republic the second spot with 796 points. The top ten countries were as follows:
Number
Country
2005 Ranking
Points
1
Argentina
4
772
2
Brazil
1
838
3
Czech Republic
2
796
4
England
9
757
5
France
=5
768
6
Holland
3
789
7
Mexico
8
766
8
Portugal
10
752
9
Spain
=5
768
10
USA
7
767
Note that the countries have been placed in alphabetical order and that at first glance this gives the impression that South American countries have performed better than European countries. An alphabetical order is used to illustrate the first column ‘number’. The ‘number’ assigned to denote countries is not in any way related to their football playing capabilities but simply serves the purpose of identification, e.g. drawing numbered balls to decide which teams may play each other in a competition. This identification number constitutes a nominal scale, which says nothing about the respective performances of the countries. So whilst England is numbered 4 and Holland is numbered 6, this does not reflect the superior performance of Holland.
A much clearer way to present the list would be to place the countries in the order of their ranking, placing Brazil at the top and Portugal at the bottom of the table. The ranking would represent an ordinal scale, where it would be clear to see that the lower the number, the better the performance. But what is still missing from the ranking is the magnitude of differences between the countries.
The only way to really understand how much one country is better than another is to examine the points awarded to each country. The points awarded out of 1000 represent an interval scale. Based on the points awarded, note that only one point separates the closely ranked France (768) and USA (767), or the USA (767) and Mexico (766), but that the difference between Brazil (838) ranked at No 1 and The Czech Republic (796) ranked at No 2 is 42 points.
Measurement and scaling
Measurement means assigning numbers or other symbols to characteristics of objects according to certain pre-specified rules.[i] We measure not the object but some characteristic of it. Thus, we do not measure consumers, only their perceptions, attitudes, preferences or other relevant characteristics. In marketing research, numbers are usually assigned for one of two reasons. First, numbers permit statistical analysis of the resulting data. Second, numbers facilitate universal communication of measurement rules and results.
Measurement
The assignment of numbers or other symbols to characteristics of objects according to certain pre-specified rules.
The most important aspect of measurement is the specification of rules for assigning numbers to the characteristics. The assignment process must be isomorphic i.e., there must be one-to-one correspondence between the numbers and the characteristics being measured. For example, the same euro (€) figures can be assigned to households with identical annual incomes. Only then can the numbers be associated with specific characteristics of the measured object, and vice versa. In addition, the rules for assigning numbers should be standardised and applied uniformly. They must not change over objects or time.
Scaling may be considered an extension of measurement. Scaling involves creating a continuum upon which measured objects are located. To illustrate, consider a scale for locating consumers according to the characteristic ‘attitude towards Formula One racing’. Each respondent is assigned a number indicating an unfavourable attitude (measured as 1), a neutral attitude (measured as 2) or a favourable attitude (measured as 3). Measurement is the actual assignment of 1, 2 or 3 to each respondent. Scaling is the process of placing the respondents on a continuum with respect to their attitude towards Formula One. In our example, scaling is the process by which respondents would be classified as having an unfavourable, neutral or positive attitude.
Scaling
The generation of a continuum upon which measured objects are located.
Primary scales of measurement
There are four primary scales of measurement: nominal, ordinal, interval and ratio.[ii] These scales are illustrated in Figure 12.1, and their properties are summarised in Table 12.1 and discussed in the following sections.
[Figure 12.1 near hear]
Table 12.1 Primary scales of measurement
Scale
Basic characteristics
Common examples
Marketing example

Permissible statistics
Descriptive
Inferential
Nominal
Numbers identify and classify objects
Student registration numbers, numbers on football players’ shirts
Gender classification, bank types
Percentages, mode
Chi-square, binomial test
Ordinal
Numbers indicate the relative positions of the objects but not the magnitude of differences between them
Rankings of the top 4 teams in the football World Cup
Ranking of service quality delivered by a number of banks. Rank order of favourite television programmes
Percentile, median
Rank-order correlation, Friedman ANOVA
Interval
Differences between objects can be compared; zero point is arbitrary
Temperature (Fahrenheit, Celsius)
Attitudes, opinions, index numbers
Range, mean, standard deviation
Product-moment correlations, t-tests, ANOVA, regression, factor analysis
Ratio
Zero point is fixed; ratios of scale values can be computed
Length, weight
Age, income, costs, sales, market shares
Geometric mean, harmonic mean
Coefficient of variation
Nominal scale
A nominal scale is a figurative labelling scheme in which the numbers serve only as labels or tags for identifying and classifying objects. For example, the numbers assigned to the respondents in a study constitute a nominal scale, thus a female respondent may be assigned a number 1 and a male respondent 2. When a nominal scale is used for the purpose of identification, there is a strict one-to-one correspondence between the numbers and the objects. Each number is assigned to only one object, and each object has only one number assigned to it.
Nominal scale
A scale whose numbers serve only as labels or tags for identifying and classifying objects with a strict one-to-one correspondence between the numbers and the objects.
Common examples include student registration numbers at their college or university and numbers assigned to football players or jockeys in a horse race. In marketing research, nominal scales are used for identifying respondents, brands, attributes, banks and other objects.
When used for classification purposes, the nominally scaled numbers serve as labels for classes or categories. For example, you might classify the control group as group 1 and the experimental group as group 2. The classes are mutually exclusive and collectively exhaustive. The objects in each class are viewed as equivalent with respect to the characteristic represented by the nominal number. All objects in the same class have the same number, and no two classes have the same number.
The numbers in a nominal scale do not reflect the amount of the characteristic possessed by the objects. For example, a high number on a football player’s shirt does not imply that the footballer is a better player than one with a low number or vice versa. The same applies to numbers assigned to classes. The only permissible operation on the numbers in a nominal scale is counting. Only a limited number of statistics, all of which are based on frequency counts, are permissible. These include percentages, mode, chi-square and binomial tests (see Chapter 18). It is not meaningful to compute an average student registration number, the average gender of respondents in a survey, or the number assigned to an average bank, as in the following example.
focus on Sports Marketing Surveys
Nominal scale
In the Racetrack study of the Formula One, the numbers 1 through to 10 were assigned to the racing teams. (see extracts from the list in Table 12.2). Thus, team 2 referred to Jaguar. It did not imply that Jaguar was in any way superior or inferior to Williams, which was assigned the number 10. Any reassignment of the numbers, such as transposing the numbers assigned to Jaguar and Williams, would have no effect on the numbering system, because the numerals did not reflect any characteristics of the teams. It is meaningful to make statements such as ‘40 per cent of French respondents named Ferrari as their favourite team’. Although the average of the assigned numbers is 5.5, it is not meaningful to state that the number of the average bank is 5.5.
Table 12.2 Illustration of primary scales of measurement
No.
Nominal scale
Ordinal scale
Interval scale
Ratio scale
€ Amount spent on merchandise on this team in the last 3 months
Sponsor
Preference rankings
Preference ratings
1–7
11–17
1
BAR
5
53
5
15
35
2
Ferrari
1
10
7
17
250
3
Jaguar
6
61
5
15
100
4
Jordan
8
82
4
14
0
5
McLaren
2
25
7
17
200
6
Minardi
9
95
4
14
0
7
Renault
3
30
6
16
100
8
Sauber
10
115
2
12
10
9
Toyota
7
79
5
15
0
10
Williams
4
45
6
16
0
Ordinal scale
An ordinal scale is a ranking scale in which numbers are assigned to objects to indicate the relative extent to which the objects possess some characteristic. An ordinal scale allows you to determine whether an object has more or less of a characteristic than some other object, but not how much more or less. Thus, an ordinal scale indicates relative position, not the magnitude of the differences between the objects. The object ranked first has more of the characteristic as compared with the object ranked second, but whether the object ranked second is a close second or a poor second is not known. Common examples of ordinal scales include quality rankings, rankings of teams in a tournament and occupational status. In marketing research, ordinal scales are used to measure relative attitudes, opinions, perceptions and preferences. Measurements of this type include ‘greater than’ or ‘less than’ judgements from the respondents.
Ordinal scale
A ranking scale in which numbers are assigned to objects to indicate the relative extent to which some characteristic is possessed. Thus, it is possible to determine whether an object has more or less of a characteristic than some other object.
In an ordinal scale, as in a nominal scale, equivalent objects receive the same rank. Any series of numbers can be assigned that preserves the ordered relationships between the objects. Ordinal scales can be transformed in any way as long as the basic ordering of the objects is maintained[iii]. In other words, any monotonic positive (order preserving) transformation of the scale is permissible, since the differences in numbers are void of any meaning other than order (see the following example). For these reasons, in addition to the counting operation allowable for nominal scale data, ordinal scales permit the use of statistics based on centiles. It is meaningful to calculate percentile, quartile, median (Chapter 18), rank-order correlation (Chapter 20) or other summary statistics from ordinal data.
focus on Sports Marketing Surveys
Ordinal scale
Table 12.2 gives a particular respondent’s preference rankings. Respondents ranked the teams in order of who they preferred, by assigning a rank 1 to the first, rank 2 to the second, and so on. Note that Ferrari (ranked 1) is preferred to McLaren (ranked 2), but how much it is preferred we do not know. Also, it is not necessary that we assign numbers from 1 to 10 to obtain a preference ranking. The second ordinal scale, which assigns a number 10 to Ferrari, 25 to McLaren and 30 to Renault, is an equivalent scale, as it was obtained by a monotonic positive transformation of the first scale. The two scales result in the same ordering of the teams according to preference.
Interval scale
In an interval scale, numerically equal distances on the scale represent equal values in the characteristic being measured. An interval scale contains all the information of an ordinal scale, but it also allows you to compare the differences between objects. The difference between any two scale values is identical to the difference between any other two adjacent values of an interval scale. There is a constant or equal interval between scale values. The difference between 1 and 2 is the same as the difference between 2 and 3, which is the same as the difference between 5 and 6. A common example in everyday life is a temperature scale. In marketing research, attitudinal data obtained from rating scales are often treated as interval data.[iv]
Interval scale
A scale in which the numbers are used to rank objects such that numerically equal distances on the scale represent equal distances in the characteristic being measured.
In an interval scale, the location of the zero point is not fixed. Both the zero point and the units of measurement are arbitrary. Hence, any positive linear transformation of the form y = a + bx will preserve the properties of the scale. Here, x is the original scale value, y is the transformed scale value, b is a positive constant, and a is any constant. Therefore, two interval scales that rate objects A, B, C and D as 1, 2, 3 and 4 or as 22, 24, 26 and 28 are equivalent. Note that the latter scale can be derived from the former by using a = 20 and b = 2 in the transforming equation.
Because the zero point is not fixed, it is not meaningful to take ratios of scale values. As can be seen, the ratio of D to B values changes from 2:1 to 7:6 when the scale is transformed. Yet, ratios of differences between scale values are permissible. In this process, the constants a and b in the transforming equation drop out in the computations. The ratio of the difference between D and B values to the difference between C and B values is 2:1 in both the scales.
Statistical techniques that may be used on interval scale data include all those that can be applied to nominal and ordinal data in addition to the arithmetic mean, standard deviation (Chapter 18), product moment correlations (Chapter 20), and other statistics commonly used in marketing research. Certain specialised statistics such as geometric mean, harmonic mean and coefficient of variation, however, are not meaningful on interval scale data. The Sports Marketing Surveys example gives a further illustration of an interval scale.
focus on Sports Marketing Surveys
Interval scale
In Table 12.2, a respondent’s preferences for the ten teams are expressed on a 7 point rating scale (where a higher number represents a greater preference for a team). We can see that although Williams received a preference rating of 6 and Sauber a rating of 2, this does not mean that Williams is preferred three times as much as Sauber. When the ratings are transformed to an equivalent 11-to-17 scale (next column), the ratings for those teams become 16 and 12, and the ratio is no longer 3 to 1. In contrast, the ratios of preference differences are identical on the two scales. The ratio of preference difference between Ferrari and Sauber to the preference difference between BAR and Sauber is 5 to 3 on both the scales.
Ratio scale
A ratio scale possesses all the properties of the nominal, ordinal and interval scales, and, in addition, an absolute zero point. Thus, in ratio scales we can identify or classify objects, rank the objects, and compare intervals or differences. It is also meaningful to compute ratios of scale values. Not only is the difference between 2 and 5 the same as the difference between 14 and 17, but also 14 is seven times as large as 2 in an absolute sense. Common examples of ratio scales include height, weight, age and money. In marketing, sales, costs, market share and number of customers are variables measured on a ratio scale.
Ratio scale
The highest scale. This scale allows the researcher to identify or classify objects, rank order the objects, and compare intervals or differences. It is also meaningful to compute ratios of scale values.
Ratio scales allow only proportionate transformations of the form y = bx, where b is a positive constant. One cannot add an arbitrary constant, as in the case of an interval scale. An example of this transformation is provided by the conversion of metres to yards (b = 1.094). The comparisons between the objects are identical whether made in metres or yards.
All statistical techniques can be applied to ratio data. These include specialised statistics such as geometric mean, harmonic mean and coefficient of variation. The ratio scale is further illustrated in the context of the Sports Marketing Surveys example.
focus on Sports Marketing Surveys
Ratio scale
In the ratio scale illustrated in Table 12.2, a respondent is asked to indicate how much they had spent on team merchandise in the last three months. Note that this respondent spent €200 on Ferrari merchandise and only €10 on Sauber. They spent 20 times more issues on Ferrari compared to Sauber. Also, the zero point is fixed because 0 means that the respondent did not spend any money on teams such as Jordan and Minardi. Multiplying these numbers by 100 to convert Euros to Cents results in an equivalent scale.
The four primary scales discussed above do not exhaust the measurement level categories. It is possible to construct a nominal scale that provides partial information on order (the partially ordered scale). Likewise, an ordinal scale can convey partial information on distance, as in the case of an ordered metric scale. A discussion of these scales is beyond the scope of this text.[v]
Metric scale
A scale that is either interval or ratio in nature.
A comparison of scaling techniques
The scaling techniques commonly employed in marketing research can be classified into comparative and non-comparative scales (see Figure 12.2).
Comparative scales involve the direct comparison of stimulus objects. For example, respondents may be asked whether they prefer Coke or Pepsi. Comparative scale data must be interpreted in relative terms and have only ordinal or rank order properties. For this reason, comparative scaling is also referred to as non-metric scaling. As shown in Figure 12.2, comparative scales include paired comparisons, rank order, constant sum scales, Q-sort and other procedures.
Comparative scales
One of two types of scaling techniques in which there is direct comparison of stimulus objects with one another.
Non-metric scale
A scale that is either nominal or ordinal in nature.
[Figure 12.2 near hear]
The major benefit of comparative scaling is that small differences between stimulus objects can be detected. As they compare the stimulus objects, respondents are forced to choose between them. In addition, respondents approach the rating task from the same known reference points. Consequently, comparative scales are easily understood and can be applied easily. Other advantages of these scales are that they involve fewer theoretical assumptions, and they also tend to reduce halo or carryover effects from one judgement to another.[vi] The major disadvantages of comparative scales include the ordinal nature of the data and the inability to generalise beyond the stimulus objects scaled. For instance, to compare Virgin Cola with Coke and Pepsi the researcher would have to do a new study. These disadvantages are substantially overcome by the non-comparative scaling techniques.
Carryover effects
Where the evaluation of a particular scaled item significantly affects the respondent’s judgement of subsequent scaled items.
In non-comparative scales, also referred to as monadic or metric scales, each object is scaled independently of the others in the stimulus set. The resulting data are generally assumed to be interval or ratio scaled.[vii] For example, respondents may be asked to evaluate Coke on a 1 to 6 preference scale (1 = not at all preferred, 6 = greatly preferred). Similar evaluations would be obtained for Pepsi and Virgin Cola. As can be seen in Figure 12.2, non-comparative scales can be continuous rating or itemised rating scales. The itemised rating scales can be further classified as Likert, semantic differential or Stapel scales. Non-comparative scaling is the most widely used scaling technique in marketing research.
Non-comparative scales
One of two types of scaling techniques in which each stimulus object is scaled independently of the other objects in the stimulus set. Also called monadic scale.
Comparative scaling techniques
Paired comparison scaling
As its name implies, in paired comparison scaling a respondent is presented with two objects and asked to select one according to some criterion.[viii] The data obtained are ordinal in nature. A respondent may state that he or she prefers Belgian chocolate to Swiss, likes Kellogg’s cereals better than supermarket home brands, or likes Adidas more than Nike. Paired comparison scales are frequently used when the stimulus objects are physical products. Coca-Cola is reported to have conducted more than 190,000 paired comparisons before introducing New Coke.[ix] Paired comparison scaling is the most widely used comparative scaling technique.
Paired comparison scaling
A comparative scaling technique in which a respondent is presented with two objects at a time and asked to select one object in the pair according to some criterion. The data obtained are ordinal in nature.
Figure 12.3 shows paired comparison data obtained to assess a respondent’s bottled beer preferences. As can be seen, this respondent made 10 comparisons to evaluate five brands. In general, with n brands, [n(n – 1)/2] paired comparisons include all possible pairings of objects.[x]
Paired comparison data can be analysed in several ways.[xi] The researcher can calculate the percentage of respondents who prefer one stimulus over another by summing the matrices of Figure 12.3 for all the respondents, dividing the sum by the number of respondents, and multiplying by 100. Simultaneous evaluation of all the stimulus objects is also possible. Under the assumption of transitivity, it is possible to convert paired comparison data to a rank order.
Transitivity of preference implies that if brand A is preferred to B, and brand B is preferred to C, then brand A is preferred to C. To arrive at a rank order, the researcher determines the number of times each brand is preferred by summing the column entries in Figure 12.3. Therefore, this respondent’s order of preference, from most to least preferred, is Carlsberg, Holsten, Stella Artois, Budvar and Grolsch. It is also possible to derive an interval scale from paired comparison data using the Thurstone case V procedure. Refer to the appropriate literature for a discussion of this procedure.[xii]
Transitivity of preference
An assumption made to convert paired comparison data to rank order data. It implies that if brand A is preferred to brand B, and brand B is preferred to brand C, then brand A is preferred to brand C.
Several modifications of the paired comparison technique have been suggested. One involves the inclusion of a neutral/no difference/no opinion response. Another extension is graded paired comparisons. In this method, respondents are asked which brand in the pair is preferred and how much it is preferred. The degree of preference may be expressed by how much more the respondent is willing to pay for the preferred brand. The resulting scale is a euro metric scale. Another modification of paired comparison scaling is widely used in obtaining similarity judgements in multidimensional scaling (see Chapter 24).
Paired comparison scaling is useful when the number of brands is limited, since it requires direct comparison and overt choice. With a large number of brands, however, the number of comparisons becomes unwieldy. Other disadvantages are that violations of the assumption of transitivity may occur, and the order in which the objects are presented may bias the results.[xiii] Paired comparisons bear little resemblance to the marketplace situation, which involves selection from multiple alternatives. Also respondents may prefer one object over certain others, but they may not like it in an absolute sense.
example
Paired comparison scaling[xiv]
The most common method of taste testing is paired comparison. Respondents are asked to taste two different prodicts and say which one they prefer. The test is done in private, either in respondent homes or some other location such as a hotel suite near to a shopping mall. In these tests a minimum of 1,000 responses is considered an adequate sample.
Ocean Spray (www.oceanspray.com) the producer of bottled and canned juices/juice drinks, makes extensive use of taste tests in developing new products. Respondents are asked to sample their new drinks which are presented in pairs. They are evaluated on taste and aspects of flavour and then choose the one they like more than the other. Taste tests showed that a segment of consumers preferred white cranberries to the strong tart taste of red cranberries. Therefore in early 2002, Ocean Spray added White Cranberry drinks, made with natural white cranberries harvested a few weeks earlier than the red variety, and Juice Spritzers, lightly carbonated juice drinks, to its product line.

[Figure 12.3 near hear]
Rank order scaling
After paired comparisons, the most popular comparative scaling technique is rank order scaling. In rank order scaling respondents are presented with several objects simultaneously and asked to order or rank them according to some criterion. For example, respondents may be asked to rank brands of cars according to overall preference. As shown in Figure 12.4, these rankings are typically obtained by asking the respondents to assign a rank of 1 to the most preferred brand, 2 to the second most preferred, and so on, until a rank of n is assigned to the least preferred brand. Like paired comparison, this approach is also comparative in nature, and it is possible that the respondent may dislike the brand ranked 1 in an absolute sense. Furthermore, rank order scaling also results in ordinal data. See Table 12.2, which uses rank order scaling to derive an ordinal scale.
Rank order scaling
A comparative scaling technique in which respondents are presented with several objects simultaneously and asked to order or rank them according to some criterion.
Rank order scaling is commonly used to measure attributes of products and services as well as preferences for brands. Rank order data are frequently obtained from respondents in conjoint analysis (see Chapter 24), since rank order scaling forces the respondent to discriminate among the stimulus objects. Moreover, compared with paired comparisons, this type of scaling process more closely resembles the shopping environment. It also takes less time and eliminates intransitive responses. If there are n stimulus objects, only (n – 1) scaling decisions need be made in rank order scaling. However, in paired comparison scaling, [n(n – 1)/2] decisions would be required. Another advantage is that most respondents easily understand the instructions for ranking. The major disadvantage is that this technique produces only ordinal data.
Finally, under the assumption of transitivity, rank order data can be converted to equivalent paired comparison data, and vice versa. This point was illustrated by examining the ‘Number of times preferred’ in Figure 12.3. Hence, it is possible to derive an interval scale from rankings using the Thurstone case V procedure. Other approaches for deriving interval scales from rankings have also been suggested.[xv]
[Figure 12.4 near hear]
Constant sum scaling
In constant sum scaling, respondents allocate a constant sum of units, such as points or euros, among a set of stimulus objects with respect to some criterion. As shown in Figure 12.5, respondents may be asked to allocate 100 points to attributes of bottled beers in a way that reflects the importance they attach to each attribute. If an attribute is unimportant, the respondent assigns it zero points. If an attribute is twice as important as some other attribute, it receives twice as many points. The sum of all the points is 100. Hence the name of the scale.
Constant sum scaling
A comparative scaling technique in which respondents are required to allocate a constant sum of units such as points, euros, chits, stickers or chips among a set of stimulus objects with respect to some criterion.
The attributes are scaled by counting the points assigned to each one by all the respondents and dividing by the number of respondents. These results are presented for three groups, or segments, of respondents in Figure 12.5. Segment I attaches overwhelming importance to price. Segment II considers a high alcoholic level to be of prime importance. Segment III values bitterness, hop flavours, fragrance and the aftertaste. Such information cannot be obtained from rank order data unless they are transformed into interval data. Note that the constant sum also has an absolute zero; 10 points are twice as many as 5 points, and the difference between 5 and 2 points is the same as the difference between 57 and 54 points. For this reason, constant sum scale data are sometimes treated as metric. Although this may be appropriate in the limited context of the stimuli scaled, these results are not generalisable to other stimuli not included in the study. Hence, strictly speaking, the constant sum should be considered an ordinal scale because of its comparative nature and the resulting lack of generalisability. It can be seen that the allocation of points in Figure 12.5 is influenced by the specific attributes included in the evaluation task.
The main advantage of the constant sum scale is that it allows for fine discrimination among stimulus objects without requiring too much time. It has two primary disadvantages, however. Respondents may allocate more or fewer units than those specified. For example, a respondent may allocate 108 or 94 points. The researcher must modify such data in some way or eliminate this respondent from analysis. Another potential problem is rounding error if too few units are used. On the other hand, the use of a large number of units may be too taxing on the respondent and cause confusion and fatigue.
[Figure 12.5 near hear]
Q-sort and other procedures
Q-sort scaling was developed to discriminate among a relatively large number of objects quickly. This technique uses a rank order procedure in which objects are sorted into piles based on similarity with respect to some criterion. For example, respondents are given 100 attitude statements on individual cards and asked to place them into 11 piles, ranging from ‘most highly agreed with’ to ‘least highly agreed with’. The number of objects to be sorted should not be less than 60 nor more than 140; a reasonable range is 60 to 90 objects.[xvi] The number of objects to be placed in each pile is pre-specified, often to result in a roughly normal distribution of objects over the whole set.
Q-sort scaling
A comparative scaling technique that uses a rank order procedure to sort objects based on similarity with respect to some criterion.
Another comparative scaling technique is magnitude estimation.[xvii] In this technique, numbers are assigned to objects such that ratios between the assigned numbers reflect ratios on the specified criterion. For example, respondents may be asked to indicate whether they agree or disagree with each of a series of statements measuring attitude towards different sports. Then they assign a number between 0 to 100 to each statement to indicate the intensity of their agreement or disagreement. Providing this type of number imposes a cognitive burden on the respondents.
Another particularly useful procedure (that could be viewed as a very structured combination of observation and depth interviewing) for measuring cognitive responses or thought processes consists of verbal protocols. Respondents are asked to ‘think out loud’ and verbalise anything going through their heads while making a decision or performing a task.[xviii] The researcher says ‘If you think anything, say it aloud, no matter how trivial the thought may be.’ Even with such an explicit instruction, the respondent may be silent. At these times, the researcher will say ‘Remember to say aloud everything you are thinking.’ Everything that the respondent says is tape recorded. This record of the respondent’s verbalised thought processes is referred to as a protocol.[xix]
Verbal protocol
A technique used to understand respondents’ cognitive responses or thought processes by having them think aloud while completing a task or making a decision.
Protocols have been used to measure consumers’ cognitive responses in actual shopping trips as well as in simulated shopping environments. An interviewer accompanies the respondent and holds a microphone into which the respondent talks. Protocols, thus collected, have been used to determine the attributes and cues used in making purchase decisions, product usage behaviour, and the impact of the shopping environment on consumer decisions. Protocol analysis has also been employed to measure consumer response to advertising. Immediately after seeing an ad, the respondent is asked to list all the thoughts that came to mind while watching the ad. The respondent is given a limited amount of time to list the thoughts so as to minimise the probability of collecting thoughts generated after, rather than during, the message. After the protocol has been collected, the individual’s thoughts or cognitive responses can be coded into three categories as illustrated in Table 12.3.[xx]
Table 12.3 Coded verbal protocols
Category
Definition
Example
Support argument
Support the claim made by the message
‘Diet Coke tastes great’
Counter-argument
Refute the claim made by the message
‘Diet Coke has an aftertaste’
Source derogation
Negative opinion about the source of the message
‘Coca-Cola is not an honest company’
Protocols are, typically, incomplete. The respondent has many thoughts that she or he cannot or will not verbalise. The researcher must take the incomplete record and infer from it a measure of the underlying cognitive response.
Non-comparative scaling techniques
Respondents using a non-comparative scale employ whatever rating standard seems appropriate to them. They do not compare the object being rated either with another object or to some specified standard, such as ‘your ideal brand’. They evaluate only one object at a time; thus, non-comparative scales are often referred to as monadic scales. Non-comparative techniques consist of continuous and itemised rating scales, which are described in Table 12.4 and discussed in the following sections.
Table 12.4 Basic non-comparative scales
Scale
Basic characteristics
Examples
Advantages
Disadvantages
Continuous rating scale
Place a mark on a continuous line
Reaction to TV commercials
Easy to construct
Scoring can be cumbersome unless computerised
Itemised rating scales
Likert scale
Degree of agreement on a 1 (strongly disagree) to 5 (strongly agree) scale
Measurement of attitudes
Easy to construct, administer and understand
More time-consuming
Semantic differential scale
Seven-point scale with bipolar labels
Brand product and company images
Versatile
Controversy as to whether the data are interval
Stapel scale
Unipolar 10-point scale, –5 to +5, without a neutral point (zero)
Measurement of attitudes and images
Easy to construct, administered over phone
Confusing and difficult to apply
Continuous rating scale
In a continuous rating scale, also referred to as a graphic rating scale, respondents rate the objects by placing a mark at the appropriate position on a line that runs from one extreme of the criterion variable to the other. Thus, the respondents are not restricted to selecting from marks previously set by the researcher. The form of the continuous scale may vary considerably. For example, the line may be vertical or horizontal; scale points, in the form of numbers or brief descriptions, may be provided; and if provided, the scale points may be few or many. Three versions of a continuous rating scale are illustrated in Figure 12.6.
Continuous rating scale
A measurement scale that has respondents rate the objects by placing a mark at the appropriate position on a line that runs from one extreme of the criterion variable to the other. The form may vary considerably. Also called graphic rating scale.
Once the respondent has provided the ratings, the researcher divides the line into as many categories as desired and assigns scores based on the categories into which the ratings fall. In Figure 12.6, the respondent exhibits a favourable attitude towards Dresdner. These scores are typically treated as interval data. The advantage of continuous scales is that they are easy to construct; however, scoring is cumbersome and unreliable.[xxi] Moreover, continuous scales provide little new information. Hence, their use in marketing research has been limited. Recently, however, with the increased popularity of computer-assisted personal interviewing and other technologies, their use has become more frequent.
[Figure 12.6 near hear]
Itemised rating scales
In an itemised rating scale, respondents are provided with a scale that has a number or brief description associated with each category. The categories are ordered in terms of scale position; and the respondents are required to select the specified category that best describes the object being rated. Itemised rating scales are widely used in marketing research and form the basic components of more complex scales, such as multi-item rating scales. We first describe the commonly used itemised rating scales – the Likert, semantic differential and Stapel scales – and then examine the major issues surrounding the use of itemised rating scales.
Itemised rating scale
A measurement scale having numbers or brief descriptions associated with each category. The categories are ordered in terms of scale position.
Likert scale
Named after its developer, Rensis Likert, the Likert scale is a widely used rating scale that requires the respondents to indicate a degree of agreement or disagreement with each of a series of statements about the stimulus objects.[xxii] Typically, each scale item has five response categories, ranging from ‘strongly disagree’ to ‘strongly agree’. We illustrate with a Likert scale for evaluating attitudes towards Renault Cars.
Likert scale
A measurement scale with five response categories ranging from ‘strongly disagree’ to ‘strongly agree’ that requires respondents to indicate a degree of agreement or disagreement with each of a series of statements related to the stimulus objects.
To conduct the analysis, each statement is assigned a numerical score, ranging either from –2 to +2 or from 1 to 5. The analysis can be conducted on an item-by-item basis (profile analysis), or a total (summated) score can be calculated for each respondent by summing across items. Suppose that the Likert scale in Figure 12.7 was used to measure attitudes towards Renault as well as Ford. Profile analysis would involve comparing the two car manufacturers in terms of the average respondent ratings for each item. The summated approach is most frequently used, and as a result, the Likert scale is also referred to as a summated scale.[xxiii] When using this approach to determine the total score for each respondent on each car manufacturer, it is important to use a consistent scoring procedure so that a high (or low) score consistently reflects a favourable response. This requires that the categories assigned to the negative statements by the respondents be scored by reversing the scale. Note that for a negative statement, an agreement reflects an unfavourable response, whereas for a positive statement, agreement represents a favourable response. Accordingly, a ‘strongly agree’ response to a favourable statement and a ‘strongly disagree’ response to an unfavourable statement would both receive scores of 5.[xxiv] In the example in Figure 12.7, if a higher score is to denote a more favourable attitude, the scoring of items 2, 4, 5 and 7 will be reversed. The respondent to this set of statements has an attitude score of 26. Each respondent’s total score for each car manufacturer is calculated. A respondent will have the most favourable attitude towards a car manufacturer with the highest score. The procedure for developing summated Likert scales is described later in the section on the development and evaluation of scales.
The Likert scale has several advantages. It is easy to construct and administer, and respondents readily understand how to use the scale, making it suitable for Internet surveys, mail, telephone or personal interviews. The major disadvantage of the Likert scale is that it takes longer to complete than other itemised rating scales because respondents have to read and fully reflect upon each statement.
Semantic differential scale
The semantic differential is a seven-point rating scale with end points associated with bipolar labels that have semantic meaning. In a typical application, respondents rate objects on a number of itemised, seven-point rating scales bounded at each end by one of two bipolar adjectives, such as ‘cold’ and ‘warm’.[xxv] We illustrate this scale in Figure 12.8 by presenting a respondent’s evaluation of Formula One racing on five attributes.
Semantic differential
A seven-point rating scale with end points associated with bipolar labels.
The respondents mark the blank that best indicates how they would describe the object being rated.[xxvi] Thus, in our example, Formula One is evaluated as exciting, innovative, safe, dynamic though uninspiring. The negative adjective or phrase sometimes appears at the left side of the scale and sometimes at the right. This controls the tendency of some respondents, particularly those with very positive or very negative attitudes, to mark the right- or left-hand sides without reading the labels.
[Figure 12.7 near hear]
[Figure 12.8 near hear]
Individual items on a semantic differential scale may be scored either on a –3 to +3 or on a 1 to 7 scale. The resulting data are commonly analysed through profile analysis. In profile analysis, means or median values on each rating scale are calculated and compared by plotting or statistical analysis. This helps determine the overall differences and similarities among the objects. To assess differences across segments of respondents, the researcher can compare mean responses of different segments. Although the mean is most often used as a summary statistic, there is some controversy as to whether the data obtained should be treated as an interval scale.[xxvii] On the other hand, in cases when the researcher requires an overall comparison of objects, such as to determine Car manufacturer preference, the individual item scores are summed to arrive at a total score.
Its versatility makes the semantic differential a popular rating scale in marketing research. It has been widely used in comparing brand, product and company images. It has also been used to develop advertising and promotion strategies and in new product development studies.[xxviii]
Stapel scale
The Stapel scale, named after its developer, Jan Stapel, is a unipolar rating scale with 10 categories numbered from –5 to +5, without a neutral point (zero).[xxix] This scale is usually presented vertically. Respondents are asked to indicate by selecting an appropriate numerical response category how accurately or inaccurately each term describes the object. The higher the number, the more accurately the term describes the object, as shown in Figure 12.9. In this example, Formula One is perceived as being prestigious but not elitist.
Stapel scale
A scale for measuring attitudes that consists of a single adjective in the middle of an even-numbered range of values.
The data obtained by using a Stapel scale can be analysed in the same way as semantic differential data. The Stapel scale produces results similar to the semantic differential.[xxx] The Stapel scale’s advantages are that it does not require a pre-test of the adjectives or phrases to ensure true bipolarity and that it can be administered over the telephone. Some researchers, however, believe the Stapel scale is confusing and difficult to apply. Of the three itemised rating scales considered, the Stapel scale is used least.[xxxi] Nonetheless, this scale merits more attention than it has received.
[Figure 12.9 near hear]
Itemised rating scale decisions
As is evident from the discussion so far, non-comparative itemised rating scales can take many different forms. The researcher must make six major decisions when constructing any of these scales:
1. The number of scale categories to use
2. Balanced versus unbalanced scale
3. Odd or even number of categories
4. Forced versus non-forced choice
5. The nature and degree of the verbal description
6. The physical form of the scale.
Number of scale categories
Two conflicting considerations are involved in deciding the number of scale categories or response options. The greater the number of scale categories, the finer the discrimination among stimulus objects that is possible. On the other hand, most respondents cannot handle more than a few categories. Traditional guidelines suggest that the appropriate number of categories should be between five and nine.[xxxii] Yet there is no single optimal number of categories. Several factors should be taken into account in deciding on the number of categories.
If the respondents are interested in the scaling task and are knowledgeable about the objects, many categories may be employed. On the other hand, if the respondents are not very knowledgeable or involved with the task, fewer categories should be used. Likewise, the nature of the objects is also relevant. Some objects do not lend themselves to fine discrimination, so a small number of categories are sufficient. Another important factor is the mode of data collection. If telephone interviews are involved, many categories may confuse the respondents. Likewise, space limitations may restrict the number of categories in mail questionnaires.
How the data are to be analysed and used should also influence the number of categories. In situations where several scale items are added together to produce a single score for each respondent, five categories are sufficient. The same is true if the researcher wishes to make broad generalisations or group comparisons. If, however, individual responses are of interest or if the data will be analysed by sophisticated statistical techniques, seven or more categories may be required. The size of the correlation coefficient, a common measure of relationship between variables (Chapter 20), is influenced by the number of scale categories. The correlation coefficient decreases with a reduction in the number of categories. This, in turn, has an impact on all statistical analysis based on the correlation coefficient.[xxxiii]
Balanced versus unbalanced scale
In a balanced scale, the number of favourable and unfavourable categories is equal; in an unbalanced scale, the categories are unequal.[xxxiv] Examples of balanced and unbalanced scales are given in Figure 12.10.
Balanced scale
A scale with an equal number of favourable and unfavourable categories.
In general, in order to obtain objective data, the scale should be balanced. If the distribution of responses is likely to be skewed, however, either positively or negatively, an unbalanced scale with more categories in the direction of skewness may be appropriate. If an unbalanced scale is used, the nature and degree of imbalance in the scale should be taken into account in data analysis.
Odd or even number of categories
With an odd number of categories, the middle scale position is generally designated as neutral or impartial. The presence, position and labelling of a neutral category can have a significant influence on the response. The Likert scale is a balanced rating scale with an odd number of categories and a neutral point.[xxxv]
The decision to use an odd or even number of categories depends on whether some of the respondents may be neutral on the response being measured. If a neutral or indifferent response is possible from at least some of the respondents, an odd number of categories should be used. If, on the other hand, the researcher wants to force a response or believes that no neutral or indifferent response exists, a rating scale with an even number of categories should be used. A related issue is whether the choice should be forced or non-forced.
Forced versus non-forced choice
On forced rating scales the respondents are forced to express an opinion because a ‘no opinion’ option is not provided. In such a case, respondents without an opinion may mark the middle scale position. If a sufficient proportion of the respondents do not have opinions on the topic, marking the middle position will distort measures of central tendency and variance. In situations where the respondents are expected to have no opinion, as opposed to simply being reluctant to disclose it, the accuracy of data may be improved by a non-forced scale that includes a ‘no opinion’ category.[xxxvi]
Forced rating scale
A rating scale that forces respondents to express an opinion because a ‘no opinion’ or ‘no knowledge’ option is not provided.
[Figure 12.10 near hear]
Nature and degree of verbal description
The nature and degree of verbal description associated with scale categories varies considerably and can have an effect on the responses. Scale categories may have verbal, numerical or even pictorial descriptions. Furthermore, the researcher must decide whether to label every scale category, label only some scale categories, or label only extreme scale categories. Surprisingly, providing a verbal description for each category may not improve the accuracy or reliability of the data. Yet, an argument can be made for labelling all or many scale categories to reduce scale ambiguity. The category descriptions should be located as close to the response categories as possible.
The strength of the adjectives used to anchor the scale may influence the distribution of the responses. With strong anchors (1 = completely disagree, 7 = completely agree), respondents are less likely to use the extreme scale categories. This results in less variable and more peaked response distributions. Weak anchors (1 = generally disagree, 7 = generally agree), in contrast, produce uniform or flat distributions. Procedures have been developed to assign values to category descriptors to result in balanced or equal interval scales.[xxxvii]
Physical form of the scale
A number of options are available with respect to scale form or configuration. Scales can be presented vertically or horizontally. Categories can be expressed by boxes, discrete lines or units on a continuum and may or may not have numbers assigned to them. If numerical values are used, they may be positive, negative or both. Several possible configurations are presented in Figure 12.11.
Two unique rating scale configurations used in marketing research are the thermometer scale and the smiling face scale. For the thermometer scale, the higher the temperature the more favourable the evaluation. Likewise, happier faces indicate evaluations that are more favourable. These scales are especially useful for children.[xxxviii] Examples of these scales are shown in Figure 12.12. Table 12.5 summarises the six decisions in designing rating scales.
[Figure 12.11 near hear]
[Figure 12.12 near hear]
Table 12.5 Summary of itemised rating scale decisions
1. Number of categories
Although there is no single, optimal number, traditional guidelines suggest that there should be between five and nine categories.
2. Balanced versus unbalanced
In general, the scale should be balanced to obtain objective data.
3. Odd or even number of categories
If a neutral or indifferent scale response is possible from at least some of the respondents, an odd number of categories should be used.
4. Forced versus unforced
In situations where the respondents are expected to have no opinion, the accuracy of the data may be improved by a non-forced scale.
5. Verbal description
An argument can be made for labelling all or many scale categories. The category descriptions should be located as close to the response categories as possible.
6. Physical form
A number of options should be tried and the best one selected.
The development and evaluation of scales
The development of multi-item rating scales requires considerable technical expertise.[xxxix] Figure 12.13 presents a sequence of operations needed to construct multi-item scales.
The characteristic to be measured is frequently called a construct. Scale development begins with an underlying theory of the construct being measured. Theory is necessary not only for constructing the scale but also for interpreting the resulting scores. The next step is to generate an initial pool of scale items. Typically, this is based on theory, analysis of secondary data and qualitative research. From this pool, a reduced set of potential scale items is generated by the judgement of the researcher and other knowledgeable individuals. Some qualitative criterion is adopted to aid their judgement. The reduced set of items may still be too large to constitute a scale. Thus, further reduction is achieved in a quantitative manner.
Data are collected on the reduced set of potential scale items from a large pre-test sample of respondents. The data are analysed using techniques such as correlations, factor analysis, cluster analysis, discriminant analysis and statistical tests discussed later in this book. As a result of these statistical analyses, several more items are eliminated, resulting in a purified scale. The purified scale is evaluated for reliability and validity by collecting more data from a different sample (these concepts will be explained on page 313). On the basis of these assessments, a final set of scale items is selected. As can be seen from Figure 12.13, the scale development process is an iterative one with several feedback loops.[xl]
A multi-item scale should be evaluated for accuracy and applicability.[xli] As shown in Figure 12.14, this involves an assessment of reliability, validity and generalisability of the scale. Approaches to assessing reliability include test–re-test reliability, alternative-forms reliability and internal consistency reliability. Validity can be assessed by examining content validity, criterion validity and construct validity.
[Figure 12.13 near hear]
[Figure 12.14 near hear]
Before we can examine reliability and validity we need an understanding of measurement accuracy; it is fundamental to scale evaluation.
Measurement accuracy
A measurement is a number that reflects some characteristic of an object. A measurement is not the true value of the characteristic of interest but rather an observation of it. A variety of factors can cause measurement error, which results in the measurement or observed score being different from the true score of the characteristic being measured (see Table 12.6).
Measurement error
The variation in the information sought by the researcher and the information generated by the measurement process employed.
Table 12.6 Potential sources of error in measurement
1 Other relatively stable characteristics of the individual that influence the test score, such as intelligence, social desirability and education
2 Short-term or transient personal factors, such as health, emotions, fatigue
3 Situational factors, such as the presence of other people, noise and distractions
4 Sampling of items included in the scale: addition, deletion or changes in the scale items
5 Lack of clarity of the scale, including the instructions or the items themselves
6 Mechanical factors, such as poor printing, overcrowding items in the questionnaire, and poor design
7 Administration of the scale, such as differences among interviewers
8 Analysis factors, such as differences in scoring and statistical analysis
The true score model provides a framework for understanding the accuracy of measurement.[xlii] According to this model,
True score model
A mathematical model that provides a framework for understanding the accuracy of measurement.

where XO = the observed score or measurement
XT = the true score of the characteristic
XS = systematic error
XR = random error
Note that the total measurement error includes the systematic error, XS, and the random error, XR. Systematic error affects the measurement in a constant way. It represents stable factors that affect the observed score in the same way each time the measurement is made, such as mechanical factors (see Table 12.6). Random error, on the other hand, is not constant. It represents transient factors that affect the observed score in different ways each time the measurement is made, such as short-term transient personal factors or situational factors (see Table 12.6). The distinction between systematic and random error is crucial to our understanding of reliability and validity.
Systematic error
An error that affects the measurement in a constant way and represents stable factors that affect the observed score in the same way each time the measurement is made.
Random error
An error that arises from random changes or differences in respondents or measurement situations.
Reliability
Reliability refers to the extent to which a scale produces consistent results if repeated measurements are made.[xliii] Systematic sources of error do not have an adverse impact on reliability, because they affect the measurement in a constant way and do not lead to inconsistency. In contrast, random error produces inconsistency, leading to lower reliability. Reliability can be defined as the extent to which measures are free from random error, XR. If XR = 0, the measure is perfectly reliable.
Reliability
The extent to which a scale produces consistent results if repeated measurements are made on the characteristic.
Reliability is assessed by determining the proportion of systematic variation in a scale. This is done by determining the association between scores obtained from different administrations of the scale. If the association is high, the scale yields consistent results and is therefore reliable. Approaches for assessing reliability include the test–re-test, alternative forms, and internal consistency methods.
In test–re-test reliability, respondents are administered identical sets of scale items at two different times, under as nearly equivalent conditions as possible. The time interval between tests or administrations is typically two to four weeks. The degree of similarity between the two measurements is determined by computing a correlation coefficient (see Chapter 20). The higher the correlation coefficient, the greater the reliability.
Test–re-test reliability
An approach for assessing reliability, in which respondents are administered identical sets of scale items at two different times, under as nearly equivalent conditions as possible.
Several problems are associated with the test–re-test approach to determining reliability. First, it is sensitive to the time interval between testing. Other things being equal, the longer the time interval, the lower the reliability. Second, the initial measurement may alter the characteristic being measured. For example, measuring respondents’ attitude towards low-alcohol beer may cause them to become more health conscious and to develop a more positive attitude towards low-alcohol beer. Third, it may be impossible to make repeated measurements (for example, the research topic may be the respondent’s initial reaction to a new product). Fourth, the first measurement may have a carryover effect to the second or subsequent measurements. Respondents may attempt to remember answers they gave the first time. Fifth, the characteristic being measured may change between measurements. For example, favourable information about an object between measurements may make a respondent’s attitude more positive. Finally, the test–re-test reliability coefficient can be inflated by the correlation of each item with itself. These correlations tend to be higher than correlations between different scale items across administrations. Hence, it is possible to have high test–re-test correlations because of the high correlations between the same scale items measured at different times even though the correlations between different scale items are quite low. Because of these problems, a test–re-test approach is best applied in conjunction with other approaches, such as alternative-forms reliability.[xliv]
In alternative-forms reliability, two equivalent forms of the scale are constructed. The same respondents are measured at two different times, usually two to four weeks apart (e.g. by initially using Likert scaled items and then using Stapel scaled items). The scores from the administrations of the alternative scale forms are correlated to assess reliability.[xlv]
Alternative-forms reliability
An approach for assessing reliability that requires two equivalent forms of the scale to be constructed and then the same respondents to be measured at two different times.
The two forms should be equivalent with respect to content, i.e. each scale item should attempt to measure the same items. The main problems with this approach are that it is difficult, time-consuming and expensive to construct an equivalent form of the scale. In a strict sense, it is required that the alternative sets of scale items should have the same means, variances and intercorrelations. Even if these conditions are satisfied, the two forms may not be equivalent in content. Thus, a low correlation may reflect either an unreliable scale or non-equivalent forms.
Internal consistency reliability is used to assess the reliability of a summated scale where several items are summed to form a total score. In a scale of this type, each item measures some aspect of the construct measured by the entire scale, and the items should be consistent in what they indicate about the construct. This measure of reliability focuses on the internal consistency of the set of items forming the scale.
Internal consistency reliability
An approach for assessing the internal consistency of the set of items, where several items are summated in order to form a total score for the scale.
The simplest measure of internal consistency is split-half reliability. The items on the scale are divided into two halves and the resulting half scores are correlated. High correlations between the halves indicate high internal consistency. The scale items can be split into halves based on odd- and even-numbered items or randomly. The problem is that the results will depend on how the scale items are split. A popular approach to overcoming this problem is to use the coefficient alpha.
Split-half reliability
A form of internal consistency reliability in which the items constituting the scale are divided into two halves and the resulting half scores are correlated.
The coefficient alpha, or Cronbach’s alpha, is the average of all possible split-half coefficients resulting from different ways of splitting the scale items.[xlvi] This coefficient varies from 0 to 1, and a value of 0.6 or less generally indicates unsatisfactory internal consistency reliability. An important property of coefficient alpha is that its value tends to increase with an increase in the number of scale items. Therefore, coefficient alpha may be artificially, and inappropriately, inflated by including several redundant scale items.[xlvii] Another coefficient that can be employed in conjunction with coefficient alpha is coefficient beta. Coefficient beta assists in determining whether the averaging process used in calculating coefficient alpha is masking any inconsistent items.
Coefficient alpha
A measure of internal consistency reliability that is the average of all possible split-half coefficients resulting from different splittings of the scale items.
Some multi-item scales include several sets of items designed to measure different aspects of a multidimensional construct. For example, car manufacturer image is a multidimensional construct that includes country of origin, range of cars, quality of cars, car performance, service of car dealers, credit terms, dealer location, and physical layout of dealerships. Hence, a scale designed to measure car manufacturer image could contain items measuring each of these dimensions. Because these dimensions are somewhat independent, a measure of internal consistency computed across dimensions would be inappropriate. If several items are used to measure each dimension, however, internal consistency reliability can be computed for each dimension.
Validity
The validity of a scale may be considered as the extent to which differences in observed scale scores reflect true differences among objects on the characteristic being measured, rather than systematic or random error. Perfect validity requires that there be no measurement error (XO = XT, XR = 0, XS = 0). Researchers may assess content validity, criterion validity or construct validity.[xlviii]
Validity
The extent to which a measurement represents characteristics that exist in the phenomenon under investigation.
Content validity, sometimes called face validity, is a subjective but systematic evaluation of how well the content of a scale represents the measurement task at hand. The researcher or someone else examines whether the scale items adequately cover the entire domain of the construct being measured. Thus, a scale designed to measure car manufacturer image would be considered inadequate if it omitted any of the major dimensions (country of origin, range of cars, quality of cars, car performance, etc.). Given its subjective nature, content validity alone is not a sufficient measure of the validity of a scale, yet it aids in a common-sense interpretation of the scale scores. A more formal evaluation can be obtained by examining criterion validity.
Content validity
A type of validity, sometimes called face validity, that consists of a subjective but systematic evaluation of the representativeness of the content of a scale for the measuring task at hand.
Criterion validity reflects whether a scale performs as expected in relation to other selected variables (criterion variables) as meaningful criteria. If, for example, a scale is designed to measure loyalty in customers, criterion validity might be determined by comparing the results generated by this scale with results generated by observing the extent of repeat purchasing. Based on the time period involved, criterion validity can take two forms, concurrent validity and predictive validity.
Criterion validity
A type of validity that examines whether the measurement scale performs as expected in relation to other selected variables as meaningful criteria.
Concurrent validity is assessed when the data on the scale being evaluated (e.g. loyalty scale) and the criterion variables (e.g. repeat purchasing) are collected at the same time. The scale being developed and the alternative means of encapsulating the criterion variables would be administered simultaneously and the results compared.
Concurrent validity
A type of validity that is assessed when the data on the scale being evaluated and on the criterion variables are collected at the same time.
Predictive validity is concerned with how well a scale can forecast a future criterion. To assess predictive validity, the researcher collects data on the scale at one point in time and data on the criterion variables at a future time. For example, attitudes towards how loyal customers feel to a particular brand could be used to predict future repeat purchases of that brand. The predicted and actual purchases are compared to assess the predictive validity of the attitudinal scale.
Predictive validity
A type of validity that is concerned with how well a scale can forecast a future criterion.
Construct validity addresses the question of what construct or characteristic the scale is, in fact, measuring. When assessing construct validity, the researcher attempts to answer theoretical questions about why the scale works and what deductions can be made concerning the underlying theory. Thus, construct validity requires a sound theory of the nature of the construct being measured and how it relates to other constructs. Construct validity is the most sophisticated and difficult type of validity to establish. As Figure 12.14 shows, construct validity includes convergent, discriminant and nomological validity.
Construct validity
A type of validity that addresses the question of what construct or characteristic the scale is measuring. An attempt is made to answer theoretical questions of why a scale works and what deductions can be made concerning the theory underlying the scale.
Convergent validity is the extent to which the scale correlates positively with other measurements of the same construct. It is not necessary that all these measurements be obtained by using conventional scaling techniques. Discriminant validity is the extent to which a measure does not correlate with other constructs from which it is supposed to differ. It involves demonstrating a lack of correlation among differing constructs. Nomological validity is the extent to which the scale correlates in theoretically predicted ways with measures of different but related constructs. A theoretical model is formulated that leads to further deductions, tests and inferences.
Convergent validity
A measure of construct validity that measures the extent to which the scale correlates positively with other measures of the same construct.
Discriminant validity
A type of construct validity that assesses the extent to which a measure does not correlate with other constructs from which it is supposed to differ.
Nomological validity
A type of validity that assesses the relationship between theoretical constructs. It seeks to confirm significant correlations between the constructs as predicted by a theory.
An example of construct validity can be evaluated in the following example. A researcher seeks to provide evidence of construct validity in a multi-item scale, designed to measure the concept of ‘self-image’. These findings would be sought:[xlix]
§ High correlations with other scales designed to measure self-concepts and with reported classifications by friends (convergent validity)
§ Low correlations with unrelated constructs of brand loyalty and variety-seeking (discriminant validity)
§ Brands that are congruent with the individual’s self-concept are more preferred, as postulated by the theory (nomological validity)
§ A high level of reliability.
Note that a high level of reliability was included as evidence of construct validity in this example. This illustrates the relationship between reliability and validity.
Relationship between reliability and validity
The relationship between reliability and validity can be understood in terms of the true score model. If a measure is perfectly valid, it is also perfectly reliable. In this case, XO = XT, XR = 0, and XS = 0. Thus, perfect validity implies perfect reliability. If a measure is unreliable, it cannot be perfectly valid, since at a minimum XO = XT + XR. Furthermore, systematic error may also be present, that is, XS ¹ 0. Thus, unreliability implies invalidity. If a measure is perfectly reliable, it may or may not be perfectly valid, because systematic error may still be present (XO = XT + XS). In other words, a reliable scale can be constructed to measure ‘customer loyalty’ but it may not necessarily be a valid measurement of ‘customer loyalty’. Conversely, a valid measurement of ‘customer loyalty’ has to be reliable. Reliability is a necessary, but not sufficient, condition for validity.
Generalisability
Generalisability refers to the extent to which one can generalise from the observations at hand to a universe of generalisations. The set of all conditions of measurement over which the investigator wishes to generalise is the universe of generalisation. These conditions may include items, interviewers, and situations of observation. A researcher may wish to generalise a scale developed for use in personal interviews to other modes of data collection, such as mail and telephone interviews. Likewise, one may wish to generalise from a sample of items to the universe of items, from a sample of times of measurement to the universe of times of measurement, from a sample of observers to a universe of observers, and so on.[l]
Generalisability
The degree to which a study based on a sample applies to the population as a whole.
In generalisability studies, measurement procedures are designed to investigate each universe of interest by sampling conditions of measurement from each of them. For each universe of interest, an aspect of measurement called a facet is included in the study. Traditional reliability methods can be viewed as single-facet generalisability studies. A test–re-test correlation is concerned with whether scores obtained from a measurement scale are generalisable to the universe scores across all times of possible measurement. Even if the test–re-test correlation is high, nothing can be said about the generalisability of the scale to other universes. To generalise to other universes, generalisability theory procedures must be employed.
Choosing a scaling technique
In addition to theoretical considerations and evaluation of reliability and validity, certain practical factors should be considered in selecting scaling techniques for a particular marketing research problem.[li] Selecting an appropriate rating scale is a necessary first step in developing a good measurement instrument, establishing statistical reliability and validity through a multi-step testing and re-testing process should be accorded the highest priority in selecting a scale. A good rating scale should have the following characteristics[lii]:
· Minimal response bias.
· Respondent interpretation and understanding
· Discriminating power
· Ease of administration
· Ease of use by respondents
· Credibility and usefulness of results

As a general rule, using the scaling technique that will yield the highest level of information feasible in a given situation will permit using the greatest variety of statistical analyses. Also, regardless of the type of scale used, whenever feasible, several scale items should measure the characteristic of interest. This provides more accurate measurement than a single-item scale. In many situations, it is desirable to use more than one scaling technique or to obtain additional measures using mathematically derived scales.
Mathematically derived scales
All the scaling techniques discussed in this chapter require the respondents to directly evaluate the constructs that the researcher believes to comprise the object of study, e.g. the cognitive state of customer satisfaction. In contrast, mathematical scaling techniques allow researchers to infer respondents’ evaluations of the constructs of the object of study. These evaluations are inferred from the respondents’ overall judgements. Two popular mathematically derived scaling techniques are multidimensional scaling and conjoint analysis, which are discussed in detail in Chapter 24.
International marketing research
In designing the scale or response format, respondents’ educational or literacy levels should be taken into account.[liii] One approach is to develop scales that are pan-cultural, or free of cultural biases. Of the scaling techniques we have considered, the semantic differential scale may be said to be pan-cultural. It has been tested in a number of countries and has consistently produced similar results. The consistency of results occurred in the following example where Xerox successfully used a Russian translation of an equivalent English semantic differential scale.
example
Copying the name Xerox[liv]
Xerox was a name well received in the former Soviet Union since the late 1960s. In fact, the act of copying documents was called Xeroxing, a term coined after the name of the company. It was a brand name people equated with quality. With the disintegration of the Soviet Union into the Commonwealth of Independent States, however, Xerox’s sales started to fall. The management initially considered this problem to be the intense competition with strong competitors such as Canon, Ricoh, Mitsubishi and Minolta. First attempts to make the product more competitive did not help. Subsequently, marketing research was undertaken to measure the image of Xerox and its competitors in Russia. Semantic differential scales were used, as examples of this type of scale translated well in other countries and were thus considered pan-cultural. The bipolar labels used were carefully tested to ensure that they had the intended semantic meaning in the Russian context.
The results of the study revealed that the real problem was a growing negative perception of Russian customers toward Xerox products. What could have wrong? The problem was not with Xerox, but with several independent producers of copying machines that had illegally infringed on Werox’s trademark rights. With the disintegration of the Soviet Union, the protection of these trademarks was unclear and trademark infringement kept growing. As a result, customers developed a misconception that Xerox was selling low quality products.
Although the semantic differential worked well in the Russian context, an alternative approach is to develop scales that use a self-defined cultural norm as a base referent. For example, respondents may be required to indicate their own anchor point and position relative to a culture-specific stimulus set. This approach is useful for measuring attitudes that are defined relative to cultural norms (e.g. attitude towards marital roles). In developing response formats, verbal rating scales appear to be the most suitable. Even less educated respondents can readily understand and respond to verbal scales. Special attention should be devoted to determining equivalent verbal descriptors in different languages and cultures. The end points of the scale are particularly prone to different interpretations. In some cultures, 1 may be interpreted as best, whereas in others it may be interpreted as worst, regardless of how it is scaled. It is important that the scale end points and the verbal descriptors be employed in a manner consistent with the culture.
Finally, in international marketing research, it is critical to establish the equivalence of scales and measures used to obtain data from different countries. This topic is complex and is discussed in some detail in Chapter 26.
Ethics in marketing research
Researchers should not bias scales so as to slant the findings in any particular direction. This is easy to do by either biasing the wording of statements (Likert type scales), the scale descriptors, or other aspects of the scales. Consider, for example, the use of scale descriptors. The descriptors used to frame a scale can be manipulated to bias results in any direction. They can be manipulated to generate a positive view of the client’s brand or a negative view of a competitor’s brand. A researcher who wants to project the client’s brand favourably can ask respondents to indicate their opinion of the brand on several attributes using seven-point scales framed by the descriptors ‘extremely poor’ to ‘good’. Using a strongly negative descriptor with only a mildly positive one has an interesting effect. As long as the product is not the worst, respondents will be reluctant to rate the product extremely poorly. In fact, respondents who believe the product to be only mediocre will end up responding favourably. Try this yourself. How would you rate BMW cars on the following attributes?
Reliability
Horrible
1
2
3
4
5
6
7
Good
Performance
Very poor
1
2
3
4
5
6
7
Good
Quality
One of the worst
1
2
3
4
5
6
7
Good
Prestige
Very low
1
2
3
4
5
6
7
Good
Did you find yourself rating BMW cars positively? Using this same technique, a researcher can negatively bias evaluations of competitors’ products by providing a mildly negative descriptor (somewhat poor) against a strong positive descriptor (extremely good).
Thus we see how important it is to use balanced scales with comparable positive and negative descriptors. When this guide is not practised, responses are biased and should be interpreted accordingly. This concern also underscores the need to adequately establish the reliability, validity and generalisability of scales before using them in a research project. Scales that are invalid, unreliable or not generalisable to the target market provide the client with flawed results and misleading findings, thus raising serious ethical issues. The researcher has a responsibility to both the client and respondents to ensure the applicability and usefulness of the scale.
Internet and computer applications
All the primary scales of measurement that we have considered can be implemented on the Internet. The same is true for the commonly used comparative scales. Paired comparisons involving verbal, visual or auditory comparisons can be implemented with ease. However, taste, smell and touch comparisons are difficult to implement. It may also be difficult to implement specialised scales such as the Q-sort. The process of implementing comparative scales may be facilitated by searching the Internet for similar scales that have been implemented by other researchers.
Continuous rating scales may be easily implemented on the Internet. The cursor can be moved on the screen in a continuous fashion to select the exact position on the scale that best describes the respondent’s evaluation. Moreover, the scale values can be automatically scored by the computer, thus increasing the speed and accuracy of processing the data.
Similarly, it is also easy to implement all of the three itemised rating scales on the Internet. Again, you can use the Internet to search for and locate cases and examples where scales have been used by other researchers. It is also possible that other researchers have reported reliability and validity assessments for multi-item scales. Before generating new scales, a researcher should first examine similar scales used by other researchers and consider using them if they meet their measurement objectives. The following example illustrates how Domino Pizza uses the Internet to conduct customer surveys and uses the full range of scale types.
example
Primary scales help Domino’s to become a primary competitor[lv]
Domino’s Pizza builds websites to communicate its image and give information on its products. It also sees its website as a medium to collect information on customers and therefore conduct marketing research. Although no pizza is sold online, the company has a main website (www.dominos.com) in addition to websites for its local subsidiaries. For local subsidiaries, the customer is asked to fill in a comment form on the website. This survey helps the local team to better understand its customers’ needs and better service them. Different scales are utilised to obtain the following information.
· Name, phone number, email address (nominal scale)
· Preference for pizza restaurants in the local area (ordinal scale)
· Impressions on the service offered by Domino’s Pizza as a whole (interval scale)
· Assessments on the products and price (interval scale)
· Customer satisfaction (interval scale)
· Amount spent on pizza and fast foods (ratio scale)
This enables the company to measure customer satisfaction and to use that information for a variety of purposes, including linking it to employee salaries.

Summary
Measurement is the assignment of numbers or other symbols to characteristics of objects according to set rules. Scaling involves the generation of a continuum upon which measured objects are located. The four primary scales of measurement are nominal, ordinal, interval and ratio. Of these, the nominal scale is the most basic in that the numbers are used only for identifying or classifying objects. In the ordinal scale, the numbers indicate the relative position of the objects but not the magnitude of difference between them. The interval scale permits a comparison of the differences between the objects. Because it has an arbitrary zero point, however, it is not meaningful to calculate ratios of scale values on an interval scale. The highest level of measurement is represented by the ratio scale in which the zero point is fixed. The researcher can compute ratios of scale values using this scale. The ratio scale incorporates all the properties of the lower-level scales.
Scaling techniques can be classified as comparative or non-comparative. Comparative scaling involves a direct comparison of stimulus objects. Comparative scales include paired comparisons, rank order, constant sum and the Q-sort. The data obtained by these procedures have only ordinal properties. Verbal protocols, where the respondent is instructed to think out loud, can be used for measuring cognitive responses.
In non-comparative scaling, each object is scaled independently of the other objects in the stimulus set. The resulting data are generally assumed to be interval or ratio scaled. Non-comparative rating scales can be either continuous or itemised. The itemised rating scales are further classified as Likert, semantic differential, or Stapel scales. When using non-comparative itemised rating scales, the researcher must decide on the number of scale categories, balanced versus unbalanced scales, an odd or even number of categories, forced versus non-forced choices, the nature and degree of verbal description, and the physical form or configuration.
Multi-item scales consist of a number of rating scale items. These scales should be evaluated in terms of reliability and validity. Reliability refers to the extent to which a scale produces consistent results if repeated measurements are made. Approaches to assessing reliability include test–re-test, alternative forms and internal consistency. The validity of a measurement may be assessed by evaluating content validity, criterion validity and construct validity.
The choice of particular scaling techniques in a given situation should be based on theoretical and practical considerations. Generally, the scaling technique used should be the one that will yield the highest level of information feasible. Also, multiple measures should be obtained.
In international marketing research, special attention should be devoted to determining equivalent verbal descriptors in different languages and cultures. The misuse of scale descriptors also raises serious ethical concerns. The researcher has a responsibility to both the client and respondents to ensure the applicability and usefulness of scales.
Questions
1. What is measurement?
2. Highlight any marketing phenomena that you feel may be problematic in terms of assigning numbers to characteristics of those phenomena.
3. Describe and illustrate, with examples, the differences between a nominal and an ordinal scale.
4. What are the advantages of a ratio scale over an interval scale? Are these advantages significant?
5. What is a comparative rating scale ?
6. What is a paired comparison? What are the advantages and disadvantages of paired comparison scaling?
7. Describe the constant sum scale. How is it different from the other comparative rating scales?
8. Identify the type of scale (nominal, ordinal, interval or ratio) used in each of the following. Give reasons for your choice.
(a) I like to listen to the radio when I am revising for exams
Disagree Agree
1 2 3 4 5
(b) How old are you? _____
(c) Rank the following activities in terms of your preference by assigning a rank from 1 to 5 (1 = most preferred, 2 = second most preferred, etc.).
(i) Reading magazines
(ii) Watching television
(iii) Going to the cinema
(iv) Shopping for clothes
(v) Eating out
(d) What is your university/college registration number? _____
(e) In an average weekday, how much time do you spend doing class assignments?
(i) Less than 15 minutes
(ii) 15 to 30 minutes
(iii) 31 to 60 minutes
(iv) 61 to 120 minutes
(v) More than 120 minutes
(f) How much money did you spend last week in the Student Union Bar? _____
9. Describe the semantic differential scale and the Likert scale. For what purposes are these scales used?
10. What are the major decisions involved in constructing an itemised rating scale? How many scale categories should be used in an itemised rating scale? Why?
11. Should an odd or even number of categories be used in an itemised rating scale?
12. How does the nature and degree of verbal description affect the response to itemised rating scales?
13. What is reliability? What are the differences between test–re-test and alternative-forms reliability?
14. What is validity? What is criterion validity? How is it assessed?
15. How would you select a particular scaling technique?
Exercises
Exercises
1. You work in the marketing research department of a firm specialising in Decision Support Systems for the health care industry. Your firm would like to measure the attitudes of hospital administrators towards Decision Support Systems produced by your firm and its main competitors. The attitudes would be measured using a telephone survey. You have been asked to develop an appropriate scale for this purpose. You have also been asked to explain and justify your reasoning in constructing this scale.
2. Develop three comparative (paired comparison, rank order and constant sum) scales to measure attitude toward five popular brands of beer (e.g. Heineken, Guinness, Carlsberg, Stella and Holsten). Administer each scale to five students. No student should be administered more than one scale. Note the time it takes each student to respond. Which scale was the easiest to administer? Which scale took the shortest time?
3. Develop a constant sum scale to determine preferences for restaurants. Administer this scale to a pilot sample of 20 students to determine their preferences for some of the popular restaurants in your town or city. Based on your pilot, evaluate the efficacy of the scale items you chose, and design new scale items that could be used for a full survey.
4. Design Likert scales to measure the usefulness of Renault’s website. Visit the site at (www.renault.com) and rate it on the scales that you have developed. After your site visit, were there any aspects of usefulness that you had not considered in devising your scales, what were they and why were they not apparent before you made your site visit?
5. In a small group discuss the following issues: “A brand could receive the highest median rank on a rank order scale of all the brands considered and still have poor sales” and “It really does not matter which scaling technique you use. As long as your measure is reliable, you will get the right results”.

Figure 12.1 An illustration of primary scales of measurement
Figure 12.2 A classification of scaling techniques
Instructions
We are going to present you with ten pairs of bottled beer brands. For each pair, please indicate which of the two brands of beer in the pair you prefer.
Recording form

Holsten
Stella Artois
Grolsch
Carlsberg
Budvar
Holsten

0
0
1
0
Stella Artois
1a

0
1
0
Grolsch
1
1

1
1
Carlsberg
0
0
0

0
Budvar
1
1
0
1

Number of times preferredb
3
2
0
4
1
a 1 in a particular box means that the brand in that column was preferred over the brand in the corresponding row. A 0 means that the row brand was preferred over the column brand.
b The number of times a brand was preferred is obtained by summing the 1s in each column.
Figure 12.3 Obtaining bottled beer preferences using paired comparisons
Instructions
Rank the listed Formula One teams in order of preference. Begin by picking out the team that you like most and assign it a number 1. Then find the second most preferred team and assign it a number 2. Continue this procedure until you have ranked all the teams in order of preference. The least preferred team should be assigned a rank of 10.
No two teams should receive the same rank number.
The criterion of preference is entirely up to you. There is no right or wrong answer. Just try to be consistent.

Brand
Rank order
1
BAR

2
Ferrari

3
Jaguar

4
Jordan

5
McLaren

6
Minardi

7
Renault

8
Sauber

9
Toyota

10
Williams

Figure 12.4 Preference for Formula One using rank order scaling
Instructions
Below are eight attributes of bottled beers. Please allocate 100 points among the attributes so that your allocation reflects the relative importance you attach to each attribute. The more points an attribute receives, the more important an attribute is. If an attribute is not at all important, assign it no points. If an attribute is twice as important as some other attribute, it should receive twice as many points.
Note: the figures below represent the mean points allocated to bottled beers by three segments of a target market.
Form
MEAN POINTS ALLOCATED

Attribute
Segment I
Segment II
Segment III
1
Bitterness
8
2
17
2
Hop flavours
2
4
20
3
Fragrance
3
9
19
4
Country where brewed
9
17
4
5
Price
53
5
7
6
High alcohol level
7
60
9
7
Aftertaste
5
0
15
8
Package design
13
3
9

Sum
100
100
100
Figure 12.5 Importance of bottled beer attributes using a constant sum scale
How would you rate the quality of Michelin tyres used in Formula One racing?
Version 1
Probably the worst √ Probably the best
Version 2
Probably the worst √ Probably the best
0 10 20 30 40 50 60 70 80 90 100
Version 3
Very bad Neither good nor bad Very good
Probably the worst √ Probably the best
0 10 20 30 40 50 60 70 80 90 100

Figure 12.6 Continuous rating scale
Instructions
Listed below are different beliefs about Renault cars. Please indicate how strongly you agree or disagree with each by putting a tick next to your choice on the following scale:
1 = Strongly disagree, 2 = Disagree, 3 = Neither agree nor disagree, 4 = Agree, 5 = Strongly agree

Strongly disagree
Disagree
Neither agree nor disagree
Agree
Strongly agree
1 Renault produce high quality cars
1
2√
3
4
5
2 Renault has poor after-sales service
1
2√
3
4
5
3 I like to visit Renault dealerships
1
2
3√
4
5
4 Renault does not offer a good range of optional extras for their cars
1
2
3
4√
5
5 The credit terms at Renault dealerships are terrible
1
2
3
4√
5
6 Renault is the embodiment of European excellence in car manufacturing
1√
2
3
4
5
7 I do not like the advertising done by Renault
1
2
3
4√
5
8 Renault has an excellent selection of car types
1
2
3
4√
5
9 The price of Renault cars are fair
1
2√
3
4
5
Figure 12.7 The Likert scale
Instructions
What does Formula One racing mean to you? The following descriptive scales, bounded at each end by bipolar adjectives, summarises characteristics of the sport. Please mark X the blank that best indicates what Formula One means to you.
Form
Formula One is:
Boring :_:_:_:_:_:X:_: Exciting
Conservative:_:_:_:_:_:X:_:Innovative
Dangerous :_:_:_:_:_:_:X: Safe
Staid :_:_:_:_:_:X:_: Dynamic
Uninspiring:_:X:_:_:_:_:_: Inspirational
Figure 12.8 Semantic differential scale
Instructions
Please evaluate how accurately each word or phrase describes Formula One racing. Select a positive number for the phrases you think describe the sport accurately. The more accurately you think the phrase describes the sport, the larger the plus number you should choose. You should select a minus number for the phrases you think do not describe the sport accurately. The less accurately you think the phrase describes the sport, the larger the negative number you should choose. You can select any number from +5 for phrases you think are very accurate, to –5 for phrases you think are very inaccurate.
Form
Formula One
+5 +5
+4 X +4
+3 +3
+2 +2
+1 +1
Prestigious Elitist
–1 –1
–2 –2 X
–3 –3
–4 –4
–5 –5
Figure 12.9 The Stapel scale
Balanced scale
Clinique moisturiser for men is:
Extremely good
Very good √
Good
Bad
Very Bad
Extremely bad
Unbalanced scale
Clinique moisturiser for men is:
Extremely good
Very good √
Good
Somewhat good
Bad
Very bad
Figure 12.10 Balanced and unbalanced scales
Figure 12.11 Rating scale configurations
Figure 12.12 Some unique rating scale configurations
Figure 12.13 Development of a multi-item scale
Figure 12.14 Scale evaluation
Notes
[i] Newell, S. J., ‘The development of a scale to measure perceived corporate credibility,’ Journal of Business Reseaerch, (June 2001) 235; Gofton, K., ‘If it moves measure it’, Marketing (Marketing Technique Supplement) (4 September 1997), 17; Nunnally, J.C., Psychometric Theory, 2nd edn (New York: McGraw-Hill, 1978), 3.
[ii] Subabrata, B.B., ‘Corporate environmentalism: The construct and its measurement,’ Journal of Business Research, 55 (3) (March 2002) 177; Stevens, S., ‘Mathematics, measurement and psychophysics’, in Stevens, S. (ed.), Handbook of Experimental Psychology (New York: Wiley, 1951).
[iii] Moshkovich, H.M., ‘Ordinal judgments in multiattribute decision analysis,’ European Journal of Operational Research, 137 (3) (March 16, 2002) 625; Cook, W.D., Kress, M. and Seiford, L.M., ‘On the use of ordinal data in data envelopment analysis’, Journal of the Operational Research Society 44(2) (February 1993), 133–40; Barnard, N.R. and Ehrenberg, A.S.C., ‘Robust measures of consumer brand beliefs’, Journal of Marketing Research 27 (November 1990), 477–84; Perreault Jr, W.D. and Young, F.W., ‘Alternating least squares optimal scaling: analysis of nonmetric data in marketing research’, Journal of Marketing Research 17 (February 1980), 1–13.
[iv] Halme, M., ‘Dealing with interval scale data in data envelopment analysis,’ European Journal of Operational Research, 137 (1) (February 16, 2002) 22; and Lynn, M. and Harriss, J., ‘The desire for unique consumer products: a new individual difference scale’, Psychology and Marketing 14(6) (September 1997), 601–16.
[v] For a discussion of these scales, refer to Miller, D.C., and Salkind, N.J. Handbook of Research Design and Social Measurement 6th ed. (Thousand Oaks, CA: Sage Publications, 2002); Taiwo, A., ‘Overall evaluation rating scales: An assessment,’ International Journal of market Research, (Summer 2000) 301-311; and Coombs, C.H., ‘Theory and methods of social measurement’, in Festinger, L. and Katz, D. (eds), Research Methods in the Behavioral Sciences (New York: Holt, Rinehart & Winston, 1953).
[vi] Bastell, R.R. and Wind, Y., ‘Product development: current methods and needed developments’, Journal of the Market Research Society 8 (1980), 122–6
[vii] There is, however, some controversy regarding this issue. See Campbell, D. T. and Russo, M.J., Social Measurement (Thousand Oaks, CA: Sage, 2001); and Amoo, T., ‘Do the numeric values influence subjects’ responses to rating scales,’ Journal of International Marketing and Marketing Research, (Feb. 2001) 41; Kang, M. and Stam, A., ‘PAHAP: a pairwise aggregated hierarchical analysis of ratio-scale preferences’, Decision Sciences 25(4) (July/August 1994), 607–24
[viii] Kellogg, D.L. and Chase, R.B., ‘Constructing an empirically derived measure for customer contact’, Management Science 41(11) (November 1995), 1734–49; Corfman, K.P., ‘Comparability and comparison levels used in choices among consumer products’, Journal of Marketing Research 28 (August 1991), 368–74
[ix] Anon, ‘Competition between Coca-Cola and Pepsi to start,’ Asiainfo Daily China News (March 19, 2002) 1; Rickard, L., ‘Remembering New Coke’, Advertising Age 66(16) (17 April 1995), 6; ‘Coke’s flip-flop underscores risks of consumer taste tests’, Wall Street Journal (18 July 1985), 25
[x] It is not necessary to evaluate all possible pairs of objects, however. Procedures such as cyclic designs can significantly reduce the number of pairs evaluated. A treatment of such procedures may be found in Bemmaor, A.C. and Wagner, U., ‘A multiple-item model of paired comparisons: Seperating chance from latent performance,’ Journal of Marketing Research 37 (4) (November 2000 514-524; and Malhotra, N.K., Jain, A.K. and Pinson, C., ‘The robustness of MDS configurations in the case of incomplete data’, Journal of Marketing Research 25 (February 1988), 95–102
[xi] For an advanced application involving paired comparison data, see Bemmaor, A.C. and Wagner, U., ‘A multiple-item model of paired comparisons: Seperating chance from latent performance,’ Journal of Marketing Research 37 (4) (November 2000 514-524; and Genest, C. and Zhang, S.S., ‘A graphical analysis of ratio-scaled paired comparison data’, Management Science 42(3) (March 1996), 335–49
[xii] Campbell, D. T. and Russo, M.J., Social Measurement (Thousand Oaks, CA: Sage, 2001; Likert, R., Roslow, S. and Murphy, G., ‘A simple and reliable method of scoring the Thurstone Attitude Scales’, Personnel Psychology 46(3) (Autumn 1993), 689–90; Thurstone, L.L., The Measurement of Values (Chicago, IL: University of Chicago Press, 1959). For an application of the case V procedure, see Malhotra, N.K., ‘Marketing linen services to hospitals: a conceptual framework and an empirical investigation using Thurstone’s case V analysis’, Journal of Health Care Marketing 6 (March 1986), 43–50
[xiii] Daniles, E. and Lawford, J., ‘The effect of order in the presentation of samples in paired comparison tests’, Journal of the Market Research Society 16 (April 1974), 127–33
[xiv] Anon, ‘Cranberry juice in a can,’ Grocer, 225 (7538) (January 26, 2002) 64; and The Beverage Network; www.bevnet.com
[xv] Bottomley, P.A., ‘Testing the reliability of weight elicitation methods: Direct rating versus point allocation,’ Journal of Marketing Research, 37 (4) (November 2000) 508-513; Herman, M.W. and Koczkodaj, W.W., ‘A Monte Carlo study of pairwise comparison’, Information Processing Letters 57(1) (15 January 1996), 25–9
[xvi] Kerlinger, F., Foundations of Behavioral Research, 3rd edn (New York: Holt, Rinehart & Winston, 1973), 583–92
[xvii] Siciliano, T., ‘Magnitude estimation,’ Quirk’s Marketing Research Review (November 1999); Noel, N.M. and Nessim, H., ‘Benchmarking consumer perceptions of product quality with price: An exploration,’ Psychology & Marketing, 13 (6) (September 1996) 591-604; and Steenkamp, J-B. and Wittink, D.R., ‘The metric quality of full-profile judgments and the number of attribute levels effect in conjoint analysis,’ International Journal of Research in Marketing, 11 (3) (June 1994) 275-286.
[xviii] Hayes, J.R., ‘Issues in protocol analysis’, in Ungson, G.R. and Braunste, D.N. (eds), Decision Making: An Interdisciplinary Inquiry (Boston, MA: Kent Publishing, 1982), 61–77
[xix] For an application of verbal protocols, see Harrison, D.A., McLaughlin, M.E. and Coalter, T.M., ‘Context, cognition and common method variance: psychometric properties and verbal protocol evidence’, Organizational Behavior and Human Decision Processes 68(3) (December 1996), 246–61; Gardial, S.F., Clemons, D.S., Woodruff, R.B., Schumann, D.W. and Bums, M.J., ‘Comparing consumers’ recall of prepurchase and postpurchase product evaluation experiences’, Journal of Consumer Research 20 (March 1994), 548–60
[xx] Mick, D.G., ‘Levels of subjective comprehension in advertising processing and their relations to ad perceptions, attitudes, and memory’, Journal of Consumer Research 18 (March 1992), 411–24; Wright, P.L., ‘Cognitive processes mediating acceptance of advertising’, Journal of Marketing Research 10 (February 1973), 53–62; Wright, P.L., ‘Cognitive responses to mass media advocacy and cognitive choice processes’, in Petty, R., Ostrum, T. and Brock, T. (eds), Cognitive Responses to Persuasion (New York: McGraw-Hill, 1978)
[xxi] Murphy, I.P., ‘RAMS helps Best Western tout worldwide positioning’, Marketing News 31(1) (6 January 1996), 25
[xxii] Amoo, T. and Friedman, H.H. Friedman, ‘Overall evaluation rating scales: An assessment,’ International Journal of market Research, 42 (3) (Summer 2000) 301-310; Albaum, G., ‘The Likert scale revisited – an alternative version’, Journal of the Market Research Society 39(2) (April 1997), 331–48; Brody, C.J. and Dietz, J., ‘On the dimensionality of 2-question format Likert attitude scales’, Social Science Research 26(2) (June 1997), 197–204; Likert, R., ‘A technique for the measurement of attitudes’, Archives of Psychology 140 (1932).
[xxiii] However, when the scale is multidimensional, each dimension should be summed separately. See Stanton, J.M., ‘Issues and strategies for reducing the length of self-report scales,’ Personnel Psychology, 55 (1) (Spring 2002) 167-194; and Aaker, J.L., ‘Dimensions of brand personality’, Journal of Marketing Research 34 (August 1997), 347–56
[xxiv] Herche, J. and Engelland, B., ‘Reversed-polarity items and scale unidimensionality’, Journal of the Academy of Marketing Science 24(4) (Fall 1996), 366–74
[xxv] Sethi, R., Smith, D.C. and Whan Park, C., ‘Cross-functional product development teams, creativity and the innovativeness of new consumer products,’ Journal of Marketing Research 38 (1) (February 2001) 73-85; and Chandler, T.A. and Spies, C.J., ‘Semantic differential comparisons of attributions and dimensions among respondents from 7 nations’, Psychological Reports 79 (3 pt 1) (December 1996), 747–58
[xxvi] Miller, D.C. and Salkind, N.J., Handbook of research design and social measurement, 6th ed. (Thousand Oaks, CA: Sage, 2002) and Bearden, W.O. and Netemeyer, R.G., Handbook of Marketing Scales: Multi-Item measures for marketing and consumer behaviour research, (Thousand Oaks, CA: Sage, 1999) 456-464; and Millar, R. and Brotherton, C., ‘Measuring the effects of career interviews on young people – a preliminary study’, Psychological Reports 79 (3 pt 2) (December 1996), 1207–15
[xxvii] There is little difference in the results based on whether the data are ordinal or interval; however, see Nishisato, S., Measurement and multivariate analysis (New York: Springer-Verlag, New York, 2002); and Gaiton, J., ‘Measurement scales and statistics: resurgence of an old misconception’, Psychological Bulletin 87 (1980), 564–7
[xxviii] Ofir, C., ‘In search of negative customer feedback; The effect of expecting to evaluate on satisfaction evaluations,’ Journal of Marketing Research, (May 2001) 170-182; Reisenwitz, T.H. and Wimbush Jr, G.J., ‘Over-the-counter pharmaceuticals: exploratory research of consumer preferences toward solid oral dosage forms’, Health Marketing Quarterly 13(4) (1996), 47–61; Malhotra, S., Van Auken, S. and Lonial, S.C., ‘Adjective profiles in television copy testing’, Journal of Advertising Research (August 1981), 21–5
[xxix] Brady, M.K., ‘Performance only measurement of service quality: A replication and extension,’ Journal of Business Research, 55 (1) (January 2002) 17; and Stapel, J., ‘About 35 years of market research in the Netherlands’, Markonderzock Kwartaalschrift 2 (1969), 3–7
[xxx] Hawkins, D.I., Albaum, G. and Best, R., ‘Stapel scale or semantic differential in marketing research?’, Journal of Marketing Research 11 (August 1974), 318–22; Menezes, D. and Elbert, N.E., ‘Alternative semantic scaling formats for measuring store image: an evaluation’, Journal of Marketing Research 16 (February 1979), 80–7
[xxxi] Devellis, R.F., Scale Development: Theories and Applications (Thousand Oaks, CA: Sage, 1991); Etzel, M.J., Williams, T.G., Rogers, J.C. and Lincoln, D.J., ‘The comparability of three Stapel scale forms in a marketing setting’, in Bush, R.F. and Hunt, S.D. (eds), Marketing Theory: Philosophy of Science Perspectives (Chicago, IL: American Marketing Association, 1982), 303–6
[xxxii] Anderson, E.W., ‘Foundations of the American customer satisfaction index,’ Total Quality Management, 11 (7) (September 2000) 5869-5882; Coleman, A.M., Norris, C.E. and Peterson, C.C., ‘Comparing rating scales of different lengths – equivalence of scores from 5-point and 7-point scales’, Psychological Reports 80(2) (April 1997), 355–62; Viswanathan, M., Bergen, M. and Childers, T., ‘Does a single response category in a scale completely capture a response?’, Psychology and Marketing 13(5) (August 1996), 457–79; Cox III, E.P., ‘The optimal number of response alternatives for a scale: a review’, Journal of Marketing Research 17 (November 1980), 407–422.
[xxxiii] Dodge, Y., ‘On asymmetric properties of the correlation coefficient in the regression setting,’ The American Statistician, 55 (1) (February 2001) 51-54; Alwin, D.F., ‘Feeling thermometers versus 7-point scales – which are better’, Sociological Methods and Research 25(3) (February 1997), 318–40; Givon, M.M. and Shapira, Z., ‘Response to rating scales: a theoretical model and its application to the number of categories problem’, Journal of Marketing Research (November 1984), 410–19; Stem Jr, D.E. and Noazin, S., ‘The effects of number of objects and scale positions on graphic position scale reliability’, in Lusch, R.E. et al., 1985 AMA Educators’ Proceedings (Chicago, IL: American Marketing Association, 1985), 370–2
[xxxiv] Jones, B.S., ‘Modeling direction and intensity in semantically balanced ordinal scales: An assessment of Congressional incumbent approval,’ American Journal of Political Science, 44 (1) (January 2000) 174; Watson, D., ‘Correcting for acquiescent response bias in the absense of a balanced scale – an application to class-consciousness’, Sociological Methods and Research 21(1) (August 1992), 52–88; Schuman, H. and Presser, S., Questions and Answers in Attitude Surveys (New York: Academic Press, 1981), 179–201
[xxxv] Morrel-Samuels, P., ‘Getting the truth into workplace surveys,’ Harvard Business Review, 80 (2) (February 2002) 111; and Spagna, G.J., ‘Questionnaires: which approach do you use?’, Journal of Advertising Research (February–March 1984), 67–70
[xxxvi] McColl-Kennedy, J., ‘Measuring customer satisfaction: Why, what and how,’ Total Quality Management, 11 (7) (September 2000) 5883-5896; Hasnich, K.A., ‘The job descriptive index revisited: questions about the question mark’, Journal of Applied Psychology 77(3) (June 1992), 377–82; Schneider, K.C., ‘Uninformed response rate in survey research’, Journal of Business Research (April 1985), 153–62
[xxxvii] Amoo, T., ‘Do numeric values influence subjects’ responses to rating scales,’ Journal of International Marketing and Market Research (February 2001) 41; Gannon, K.M. and Ostrom, T.M., ‘How meaning is given to rating scales – the effects of response language on category activation’, Journal of Experimental Social Psychology 32(4) (July 1996), 337–60; Friedman, H.H. and Leefer, J.R., ‘Label versus position in rating scales’, Journal of the Academy of Marketing Science (Spring 1981), 88–92
[xxxviii] Alwin, D.F., ‘Feeling thermometers versus 7-point scales – which are better’, Sociological Methods and Research 25(3) (February 1997), 318–40
[xxxix] For an example of a multi-item scale, see Brown, T., ‘The customer orientation of service workers: Personality trait effects on self and supervisor-performance ratings,’ Journal of Marketing Research, 39 (1) (February 2002) 110-119; Mathwick, C., Malhotra, N.K. and Ridgon, E., ‘Experiential value: Conceptualization, measurement and application in the catalog and internet shopping environment,’ Journal of Retailing, 77 (2001) 39-56; Aaker, J.L., ‘Dimensions of brand personality’, Journal of Marketing Research 34 (August 1997), 347–56.
[xl] For example, see Flynn, L. R. and Pearcy, D., ‘Four subtle sins in scale development: Some suggestions for strengthening the current paradigm,’ International Journal of Market Research, 43 (4) (Fourth Quarter 2001) 409-423 and King, M.F., ‘Social desirability bias: Aneglected aspect of validity testing,’ Psychology and Marketing, 17 (2) (February 2000) 79; Singhapakdi, A., Vitell, S.J., Rallapalli, K.C. and Kraft, K.L., ‘The perceived role of ethics and social responsibility: a scale development’, Journal of Business Ethics 15(11) (November 1996), 1131–40
[xli] Borman, W.C., ‘An examination of the comparative reliability, validity and accuracy and performance ratings made using computerised adaptive rating scales,’ Journal of Applied Psychology, 86 (5) (October 2001) 965; Kim, K. and Frazier, G.L., ‘Measurement of distributor commitment in industrial channels of distribution’, Journal of Business Research 40(2) (October 1997), 139–54; Greenleaf, E.A., ‘Improving rating scale measures by detecting and correcting bias components in some response styles’, Journal of Marketing Research 29 (May 1992), 176–88
[xlii] The true score model is not the only theory of measurement. See Lord, E.M. and Novick, M.A., Statistical Theories of Mental Test-Scores (Reading, MA: Addison-Wesley, 1968)
[xliii] Thompson, B., Score reliability: Contemporary thinking on reliability issues (Thousand Oaks, CA: Sage 2002); Sinha, P., ‘Determination of reliability of estimations obtained with survey research: A method of simulation,’ International Journal of Market Research, 42 (3) (Summer 2000) 311-317; Wilson, E.J., ‘Research design effects on the reliability of rating scales in marketing – an update on Churchill and Peter’, Advances in Consumer Research 22 (1995), 360–5; Perreault Jr, W.D. and Leigh, L.E., ‘Reliability of nominal data based on qualitative judgements’, Journal of Marketing Research 25 (May 1989), 135–48; Peter, J.P., ‘Reliability: a review of psychometric basics and recent marketing practices’, Journal of Marketing Research 16 (February 1979), 6–17
[xliv] Campbell, D. T. and Russo, M.J., Social Measurement (Thousand Oaks, CA: Sage, 2001); Lam, S.S.K. and Woo, K.S., ‘Measuring service quality: a test–re-test reliability investigation of SERVQUAL’, Journal of the Market Research Society 39(2) (April 1997), 381–96
[xlv] Hunt, D., Measurement and scaling in statistics, (London: Edward Arnold, 2001); Armstrong, D., Gosling, A., Weinman, J. and Marteau, T., ‘The place of inter-rater reliability in qualitative research: an empirical study’, Sociology: The Journal of the British Sociological Association 31(3) (August 1997), 597–606; Segal, M.N., ‘Alternate form conjoint reliability’, Journal of Advertising Research 4 (1984), 31–8
[xlvi] Cronbach, L.J., ‘Coefficient alpha and the internal structure of tests’, Psychometrika 16 (1951), 297–334
[xlvii] Brown, T.J., Mowen, J.C., Donavan, D.T. and Licata, J.W., ‘The customer orientation of service workers: Personality trait effects on self- and supervisor performance ratings,’ Journal of Marketing Research, 39 (1) (February 2002) 110-119; Peterson, R.A., ‘A meta-analysis of Cronbach’s coefficient alpha’, Journal of Consumer Research 21 (September 1994), 381–91
[xlviii] Chen, G., ‘Validation of a new general self-efficacy scale,’ Organizational Research Methods, 4 (1) (January 2001) 62-83; McTavish, D.G., ‘Scale validity – a computer content analysis approach’, Social Science Computer Review 15(4) (Winter 1997), 379–93; Peter, J.P., ‘Construct validity: a review of basic issues and marketing practices’, Journal of Marketing Research 18 (May 1981), 133–45
[xlix] For further details on validity, see Keillor, ‘A cross-cultural/cross national study of influencing factors and socially desirable response biases,’ International Journal of Market Research (1st Quarter 2001) 63-84; Sirgy, M.J., Grewal, D., Mangleburg, T.F., Park, J. et al., ‘Assessing the predictive ability of two methods of measuring self-image congruence’, Journal of the Academy of Marketing Science 25(3) (Summer 1997), 229–41; Spiro, R.L. and Weitz, B.A., ‘Adaptive selling: conceptualization, measurement, and nomological validity’, Journal of Marketing Research 27 (February 1990), 61–9
[l] For a discussion of the generalisability theory and its applications in marketing research, see Middleton, K.L., ‘Socially desirable response sets: The impact of country culture,’ Psychology and Marketing, (February 2000) 149; Abe, S., Bagozzi, R.P. and Sadarangani, P., ‘An investigation of construct validity and generalizability in the self concept: self consciousness in Japan and the United States’, Journal of International Consumer Marketing 8(3,4) (1996), 97–123; Rentz, J.O., ‘Generalisability theory: a comprehensive method for assessing and improving the dependability of marketing measures’, Journal of Marketing Research 24 (February 1987), 19–28
[li] Myers, M., ‘Academic insights: An application of multiple-group causal models in assessing cross-cultural measurement equivalence,’ Journal of International Marketing, 8 (4) (2000) 108-121; Hinkin, T.R., ‘A review of scale development practices in the study of organisations’, Journal of Management 21(5) (1995), 967–88
[lii] Devlin, S.J., Dong, H.K. and Brown, M., ‘Selecting a scale for measuring quality’, Marketing Research, (Fall 2003) 13-16
[liii] Page Fisk, A., ‘Using individualism and collectivism to compare cultures- a critique of the validity and measurement of the constructs: Comment on Oyserman,’ Psychological Bulletin, 128 (1) (January 2002) 78; Mullen, M.R., Milne, G.R. and Didow, N.M., ‘Determining cross-cultural metric equivalence in survey research: a new statistical test’, Advances in International Marketing 8 (1996), 145–57; Gencturk, E., Childers, T.L. and Ruekert, R.W., ‘International marketing involvement – the construct, dimensionality, and measurement’, Journal of International Marketing 3(4) (1995), 11–37
[liv] Unikel, A.L., ‘Imitation might be flattering, but beware of trademark infringement,’ Marketing News, 21 (19) (September 11, 1997) 20021; Mckay, B., ‘Xerox fights trademark battle’, Advertising Age International (27 April 1992), 1–39.
[lv] Zuber, A., ‘Pizza chains top customer satisfaction poll,’ Nation’s Restaurant News, 36 (9) (March 4, 2002) 4-5 and www.dominos.com.

Qualitative research: focus group discussions

2010-06-05T09:04:00.000-07:00

Stage 1 Problem definition
Stage 2 Research approach developed
Stage 3 Research design developed
Stage 4 Fieldwork or data collection
Stage 5 Data preparation and analysis
Stage 6 Report preparation and presentation
Objectives
After reading this chapter, you should be able to:
1. understand why the focus group is defined as a direct qualitative research technique;
2. describe focus groups in detail, with an emphasis on planning and conducting focus groups;
3. evaluate the advantages, disadvantages and applications of focus groups;
4. describe alternative ways of conducting qualitative research in groups;
5. discuss the considerations involved in conducting qualitative research in an international setting, extending the contrast between European and US traditions of running focus groups;
6. understand the ethical issues involved in conducting focus groups;
7. describe the difference between real-time and non-real-time online focus groups.
The best moderators of focus groups are those that create a spirit of spontaneity and a passion for the issues under discussion.
Overview
In this chapter, we start by presenting a means of classifying qualitative research techniques and we examine the implications of such classification. The characteristics of the focus group are presented along with their advantages and disadvantages. The manner in which focus groups should be planned and conducted is then presented. Running successful focus groups depends upon the skills of a moderator, i.e. the person who manages the group, and the ensuing discussion. We present the qualities needed in moderators to get the most out of focus group discussions. There are variations on the main theme of running a focus group; these are described as well as other qualitative group activities. Some of the misconceptions of focus groups are examined, with a reminder of the qualities that make group techniques work well. Running focus groups using the Internet is a rapidly developing technique; we describe how focus groups can be run in ‘real-time’ and ‘non-real-time’. In Chapter 6 we contrasted the purpose and different ways of running focus groups in the US and in Europe; this contrast is developed and illustrated further in examining international marketing research issues. Several ethical issues that arise in running focus groups are identified.
The following example illustrates how using focus groups helps researchers and decision-makers to understand the issues faced by consumers and in this case, advertisement viewers. The issues are expressed in consumers’ own words and sometimes in ways that words cannot convey. The example also illustrates that researchers and decision-makers may think that they know of all the issues they should be questioning, but once the exploration starts, participants can reveal new issues that they perceive to be of more importance, and these issues can drive very innovative marketing decisions.
example
Save the Children’s emotive appeal[i]
Save the Children’s work is focused on helping the world’s most vulnerable children. A new direct response television advertisement in September 2004 formed part of a wider communications plan which included press inserts, online advertising and door-drops. In order to achieve an acceptable return-on-investment, television channels chosen to screen the advertisements were predominantly cable/satellite with low active viewing, such as MTV and Paramount Comedy. Low activity enables the viewer to walk away from what they have been watching and telephone the donation line, rather than remaining to view the rest of the programme. The advertisement began with a startling call to attention, a ‘gunshot’, to grab the viewer. Save the Children could draw upon a broad archive of relevant, harrowing footage, so focus groups were convened to assess the acceptable bounds of the imagery that could be used that would encourage an instant response. One of the key focus group findings was that, in addition to presenting ‘the problem’, the advertisement must also offer hope, reassurance that a donation could actually make a difference. For these individuals, a very emotive appeal was required to stir them into action.■

Classifying qualitative research techniques
A classification of qualitative research techniques is presented in Figure 7.1.
These techniques are classified as either direct or indirect, based on whether the true purpose of the project is known to the participants. A direct approach is not disguised. The purpose of the project is disclosed to the participants or is otherwise obvious to them from the questions asked. Focus groups and in-depth interviews are the major direct techniques. Even though the purpose of the project is disclosed, the extent to which the purpose of the research is revealed at the start of a focus group or in-depth group may vary. Suppose that the researcher wanted to understand how participants felt about the Prada brand, what their views were of Prada advertising campaigns, the style and quality of Prada clothes, how ‘cool’ the brand was, the importance of its being an Italian company – to name but a few issues that could be tackled. Rather than stating these objectives or even that the study was for Prada right at the start, the researcher may initially hide these issues. If revealed at the start, participants may focus straight on to these issues and not the surrounding contextual issues that may reveal the ‘relative’ impact of the Prada brand. Thus the researcher may initially reveal that the discussion is going to be about ‘what clothes mean to you’. The researcher may explore what participants feel to be good and poor examples of clothing advertisements and why. What types of clothing and accessories do participants see as stylish, how important is it to wear stylish clothes, how important is it to wear ‘cool’ clothes? – drawing out examples of brands to illustrate these views. Italy as a country could be explored in terms of characteristics of Italians or Italian design and style. If participants bring up Prada in the discussion, the researcher can then focus upon specific questions about the brand, contrast it with other brands and clearly see which subjects generated positive or negative views of Prada. Participants may deduce that the study is being conducted for Prada as the discussion proceeds, this may be apparent by the end of the discussion, or the researcher may clarify this point and explain why it was not revealed at the beginning.
Direct approach
A type of qualitative research in which the purposes of the project are disclosed to the participant or are obvious given the nature of the interview.
[Figure 7.1 near here]
In using focus groups or in-depth interviews, the researcher employs a direct approach but has control over how much ‘directness’ they reveal at the start of the discussion. The researcher must consider what ‘frame of mind’ they want participants to be in at the start of the discussion, as a too narrow or set focus at the start can impede the thought processes and creativity of the participants and the success of the discussion.
In contrast, research that takes an indirect approach totally disguises the purpose of the project. In an indirect approach, the researcher wants participants to behave as naturally as possible without any impediment of research purposes. In observation or ethnographic techniques, consumers may be seen shopping, choosing products, using products, interacting with other people and objects, hopefully in a natural environment and a natural manner. The ‘participant’ may not know that they are being observed, or if they do and have agreed to be observed, may not really know why. The purpose of using projective techniques (presented in Chapter 8) is to discover underlying motivations, beliefs, attitudes or feelings regarding consumer behaviour. The techniques allow indirect questioning to allow participants to discover novel ways to think about and express their feelings, where direct questioning would fail.
Indirect approach
A type of qualitative research in which the purposes of the project are disguised from the participants.
Figure 7.1 presents a useful way to remember which qualitative techniques tend towards directness and indirectness. Another way of thinking about these issues would be to visualise a continuum with ‘totally direct’ at one extreme and ‘totally indirect’ at the other. Qualitative techniques may then be positioned on this continuum and the implications of that position addressed. The implications for the researcher are as follows.
§ Ethical. What are the ethical issues concerning revealing what a study is about? Would a participant get involved in the study if they knew what it was really about? Will a participant feel cheated or abused by not being told the purpose or finding it out as they go along?
§ Data richness. If the participant knows what a study is about, to what extent does this ‘close’ their mind or destroy their creativity? Qualitative techniques aim to draw out deeply held views, issues that may be difficult to conceive or express. Researchers need to be able to get participants in the right frame of mind to be able to elicit this rich data. To what extent does revealing the purpose of a study impede this process?
The researcher cannot resolve this issue by stating, for example, that ‘they will only use direct techniques’ to resolve the ethical issues. Successful focus groups and in-depth interviews can utilise certain observation techniques. As an example, consider recording a simple answer of ‘no’ to a question. This ‘no’ may be interpreted in different ways, depending upon facial expressions (‘was the participant smiling?’), the tone of their voice (‘was it sharp and direct?’), their posture and body language (‘were they hunched and hiding their face?’), or their positioning and reactions to others around them (‘were they seeking support from others of the same view, by gestures directed towards those participants?’). The researcher can use and manipulate scenarios to observe participants as a means of interpreting the answers they give to questions. The same can be said of projective techniques, all of which can be used to great effect in focus groups and in-depth interviews.
The researcher ultimately has to work out the extent of directness or indirectness of their chosen qualitative techniques and address the ethical and data richness issues before setting out the detail of how they will administer their qualitative techniques. These issues may be unique in each investigation, depending upon the nature of participants being studied and the questions they face. In practice, the qualitative researcher may resolve the best means to administer a technique by experimenting and adapting; these issues are tackled later in this chapter.
Focus group discussions
A focus group is a discussion conducted by a trained moderator in a non-structured and natural manner with a small group of participants. A moderator leads and develops the discussion. The main purpose of focus groups is to gain insights by creating a forum where participants feel sufficiently relaxed to reflect and to portray their feelings and behaviour, at their pace and using their language and logic. It has been argued that the single most compelling purpose of the focus group is to bridge social and cultural differences between researchers and their target participants[ii]. The value of the technique and its role in bridging social and cultural differences lies in discovering unexpected findings, often obtained from a free-flowing discussion that is respectful and not condescending to participants[iii]. Focus groups are the most important qualitative marketing research procedure accounting for 11% of all global marketing research expenditure in 2004[iv]. They are used extensively in new product development, advertising development and image studies. They are so popular that many marketing research practitioners consider this technique synonymous with qualitative research.[v] Given their importance and popularity, we describe the salient characteristics of focus groups in detail.[vi]
Focus group
A discussion conducted by a trained moderator among a small group of participants in an unstructured and natural manner.
Moderator
An individual who conducts a focus group interview, by setting the purpose of the interview, questioning, probing and handling the process of discussion.
Characteristics
The major characteristics of a focus group are summarised in Table 7.1.
Table 7.1 Characteristics of focus groups
Key benefit
Group members ‘feed’ off each other and creatively reveal ideas that the researcher may not have thought of or dared to tackle
Key drawback
Group members may feel intimidated or shy and may not reveal anything
Group size
6 to 10
Group composition
Homogeneous, participants pre-screened
Physical setting
Relaxed, informal atmosphere
Stimulating discussion
Use of storyboards, mood boards, products, brochures
Time duration
1.5 to 6 hours
Recording
Use of audiocassettes, videotapes and notes from observations
Moderator
Observational, interpersonal and communication skills
One of the main characteristics and key benefits lies in the amount of creative discussion and other activities that may be generated. Group members have the time to reflect upon the discussion and range of stimuli that may be presented to them. The stimuli may come from other group members or from the moderator. Using their intuition and imagination, group members can explain how they feel or behave, in words they are comfortable with and using logic that is meaningful to them. The key drawback lies in how intimidating the group scenario may be to certain individuals. Many individuals may be self-conscious in expressing their ideas, feeling they may be ridiculed by others, or they may be shy and unable to freely express themselves in a group. A focus group is generally made up of 6–10 members. Groups of fewer than six are unlikely to generate the momentum and group dynamics necessary for a successful session. Likewise, groups of more than 10 may be too crowded and may not be conducive to a cohesive and natural discussion.[vii] Large groups have a tendency to splinter into sub-groups as group members compete to get their views across.
A focus group generally should be homogeneous in terms of demographic and socio-economic characteristics. Commonality among group members avoids inter-actions and conflicts among group members on side issues.[viii] An amount of conflict may draw out issues or get participants to rationalise and defend their views in a number of ways; it can also mean that the discussion does not get stale with everybody agreeing with each other and setting a scenario where genuine disagreement gets stifled. However, major conflicts should and can be avoided by the careful selection of participants. Thus, for many topics, a women’s group should not combine married homemakers with small children, young unmarried working women and elderly divorced or widowed women, because their lifestyles are substantially different. Participants should be carefully screened to meet stated criteria. These criteria are set by the researcher to ensure that participants have had adequate experience with the object or issue being discussed. The most fundamental basis screening participants is through demographic classification. Common demographic characteristics for determining group composition are: gender, race or ethnicity, age, household location, education level, occupation, income and marital status or family composition. Selecting participants using these characteristics can help increase compatability but it does not guarantee it, their background should be carefully balanced with their experiences[ix]. Participants who have already participated in numerous focus groups should not be included. These so-called professional participants are atypical, and their participation leads to serious validity problems.[x]
The physical setting for the focus group is also important. A relaxed, informal atmosphere helps group members to forget they are being questioned and observed. What is meant by a relaxed, informal atmosphere may change depending upon the type of participant and the subject being tackled. Examples of what ‘relaxed and informal’ means can include the home of a friend within a particular community, a works canteen, a village hall, a room in a leisure centre, a meeting room in a hotel or a purpose-built discussion group room. The poor acoustics and hard seats of a works canteen may not seem relaxed and informal. To group participants, however, it may be the place where they are happy to talk and willing to open up to a moderator. The practitioner Wendy Gordon (see Professional Perspective X on the Companion Website) contends that the issue of ‘real’ context as opposed to simulation has always been a problematic issue. She has seen an increase in the tendency to conduct groups wherever the product is seen, bought or used rather than in recruiter living rooms or purpose-built viewing facilities. An example of this could be using part of a furniture store to discuss issues around house decoration, furnishings and cleaning or maintaining the home. Such a setting may set a very strong frame of reference to start the focus group and provide lots of stimuli. This example does not mean that all research needs to be conducted in situ, but that the technique can be designed to allow the findings from the real and the research environments to inform the overall recommendations.[xi] Light refreshments should be served before the session and made available throughout; these become part of the context of relaxation. The nature of these refreshments largely depends upon how long the discussion lasts, the nature of tasks faced by the participants and the ethical viewpoint of the researcher.
Although a focus group may last from one to six hours, a duration of one and a half to two hours is typical. When a focus group lasts up to six hours, participants may be performing a series of projective techniques such as building ‘mood boards’ or ‘role playing’. Participants can be given additional stimuli to discuss such as advertising story boards and products (new, existing, competitors) to handle and examine. A focus group that lasts for up to six hours will invariably require a break for a meal, but in all circumstances a flow of drinks and snacks should be made available with note for special dietary requirements, e.g. vegan food and drinks. The opportunities and problems that occur by serving alcoholic drinks in focus groups will be discussed in the ‘ethics in marketing research’ section later in this chapter.
The lengthy period of discussion in a focus group is needed to establish rapport with the participants, to get them to relax and be in the right frame of mind, and to explore in depth their beliefs, feelings, ideas, attitudes and insights regarding the topics of concern. Focus group discussions are invariably recorded, mostly using audiotape but often on videotape, for subsequent replay, transcription and analysis. Videotaping has the advantage of recording facial expressions and body movements, but it can increase the costs significantly. Frequently, where focus groups are conducted in purpose-built studios, decision-makers as ‘clients’ observe the session from an adjacent room using a two-way mirror or through video transmission as described in the following example.
example
How wrong can you be?[xii]
Many management boards of blue-chip companies now watch consumer qualitative research as a regular part of their board activities. EVO Research and Consulting have managed this approach on many occasions. In one instance, a European board agreed that the business as a whole needed to get closer to its customers and they must lead by example. EVO organised a one-off research event in which the board would view a cross-section of their customer base taking part in a series of group discussions. To make it more interesting, the board members who were watching via a two way mirror, were each asked to make some assumptions about the types of people they were expecting to see; the lives they imagined these people led; and importantly, the types of relationships thay felt the participants had with various brands within their portfolio. The most outstanding aspect of this exercise was the level of surprise at how ‘wrong’ board perceptions of customers could be, especially when they prided themselves by how ‘in-touch’ they were with their customers.

The approach taken in the above example could work even if board members could not physically attend the group discussion. Video transmission technology also enables the clients to observe focus group sessions live from a remote location (see www.focusvision.co.uk for an example). Care must be taken with both audio and video recording in terms of how comfortable participants are with being recorded and what effects recorders have on how much they relax and honestly portray how they feel, especially when projective techniques are used (these techniques will be detailed in Chapter 8) , which some participants may find embarrassing. Many moderators can give rich examples of how the most interesting points to emerge from a focus group occur when recorders are switched off at the end of the discussion. This happens when a group of participants have really become involved in the subject of discussion and have enjoyed talking to their fellow participants. Even when the moderator has finished the discussion, some participants carry on discussing the issues between themselves as they put their coats on, leave the room and even perhaps as they walk to their cars. Moderators can hear issues discussed in this informal manner that they wish had been tackled with the full group.
The moderator plays a vital role in the success of a focus group. The moderator must establish rapport with the participants and keep the discussion flowing, including the probing of participants to elicit insights. Typically, probing differs from questioning in that they are more spontaneous, they involve comments such as[xiii]:
· Would you explain further?
· Can you give me an example of what you mean?
· Would you say more?
· Is there anything else?
· Please describe what you mean?
· I don’t understand.
· Tell me more about that.
· How does that work?
Sometimes the moderator may ask a probe question to the whole group such as:
· Who else has something?
· What about the rest of you?
· I see people nodding their heads; tell me about it.
· We want to hear all the different points of view. Who else has something that might be a bit different?
It is seen as good practice to probe early in the discussion in order to communicate the importance of precision or a developed explanation, and to use probes sparingly in later discussion.
In addition, the moderator may have a central role in the analysis and interpretation of the data. Therefore, the moderator should possess skill, experience, knowledge of the discussion topic, and an understanding of the nature of group dynamics.
Probing
A motivational technique used when asking questions to induce the participants to enlarge on, clarify or explain their answers.
Advantages and disadvantages of focus groups
Focus groups offer several advantages over other data collection techniques. These may be summarised by the 9 Ss:[xiv]
1. Synergy. Putting a group of people together will produce a wider range of information, insight and ideas than will individual responses secured privately.
2. Snowballing. A bandwagon effect often operates in a group discussion in that one person’s comment triggers a chain reaction from the other participants. This process facilitates a very creative process where new ideas can be developed, justified and critically examined.
3. Stimulation. Usually after a brief introductory period, the participants want to express their ideas and expose their feelings as the general level of excitement over the topic increases in the group.
4. Security. Because the participants’ feelings may be similar to those of other group members, they feel comfortable and are therefore willing to ‘open up’ and reveal thoughts where they may have been reluctant if they were on their own.
5. Spontaneity. Because participants are not required to answer specific questions, their responses can be spontaneous and unconventional and should therefore provide an accurate idea of their views.
6. Serendipity. Ideas are more likely to arise unexpectedly in a group than in an individual interview. There may be issues that the moderator had not thought of. The dynamics of the group can allow these issues to develop and be discussed. Group members, to great effect, may clearly and forcibly ask questions that the moderator may be reluctant to ask.
7. Specialisation. Because a number of participants are involved simultaneously, the use of a highly trained, but expensive, interviewer is justified.
8. Scientific scrutiny. The group discussion allows close scrutiny of the data collection process in that observers can witness the session and it can be recorded for later analysis. Many individuals can be involved in the validation and interpretation of the collected data.
9. Structure. The group discussion allows for flexibility in the topics covered and the depth with which they are treated. The structure can match the logical structure of issues from the participants’ perspective as well as the language and expressions they are comfortable with.

Disadvantages of focus groups may be summarised by the five Ms:
1. Misjudgement. Focus group results can be more easily misjudged than the results of other data collection techniques. As discussed in Chapter 6, as a qualitative technique, focus groups can evolve through a line of questioning and probing. The specific direction of questioning and the ultimate interpretation of findings can be susceptible to the bias of the moderator and other researchers working on a project.
2. Moderation. As well as being great fun to moderate, focus groups can be difficult to moderate. Much depends upon the ‘chemistry’ of the group in terms of how group members get on with each other and draw ideas and explanations from each other. Even moderators with many years of experience may not connect with particular groups of respondents or topics and can get into difficulty with group members who disrupt the discussion. The quality of the results depends upon how well the discussion is managed and ultimately on the skills of the moderator.
3. Messiness. The unstructured nature of the responses makes coding, analysis and interpretation difficult in comparison with the structure of quantitative techniques. Focus group data tend to be messy and need either strong theoretical support or the discipline of a grounded theory approach to ensure that decision-makers can rely upon the analyses and interpretations.
4. Misrepresentation. Focus group results concentrate on distinct target groups, describing them and contrasting them to other groups or types of participant. Trying to generalise to much wider groups, in the same manner as with a quantitative survey based on a representative sample, can be very misleading.
5. Meeting. There are many problems in getting potential participants to agree to take part in a focus group discussion. Even when they have agreed to participate, there are problems in getting focus group participants together at the same time. Running focus groups on the Internet has helped to resolve these problems to some extent, but for some target groups even this does not offer a solution. An example is in conducting business research with managers as participants. Given the amount of travel and tight schedules that many managers have, getting them together at the same time is very difficult. With many managers reluctant to reveal their company’s behaviour and plans in front of other managers, one can see that the focus group may be very difficult to administer in getting managers to meet up and discuss issues.
Planning and conducting focus groups
The procedure for planning and conducting focus groups is described in Figure 7.2.
Planning begins with an examination of the marketing research problem(s) and objectives. In most instances, the problem has been defined by this stage, but it is vital to ensure that the whole process is founded upon a clear awareness of the gaps in the knowledge of marketing decision-makers. Given the problem definition, the objectives of using focus groups should be clarified. There should be a clear understanding of what information can be elicited and what the limitations of the technique are.
[Figure 7.2 near here]
The next step is to develop a list of issues, or topic guide, that are to be tackled in the focus groups. This list may be a series of specific questions but is more likely to be a set of broad issues that can be developed into questions or probes as the focus group actually takes place. Specific questions may be of help to the moderator who feels that a consistent set of points needs to be presented to different groups in order to allow clear comparisons to be made. Specific questions also act as a ‘prop’ when the discussion is failing; indeed some group participants may initially feel that their role is to react to specific questions. However, treating the whole discussion as a means to present set questions may stifle the creativity and spontaneity that are the hallmarks of successful focus groups. The moderator should open the discussion with a general introductory question to make participants comfortable with the subject and the purpose of the research. This question should encourage conversation and interaction amongst the participants. It should not be threatening, it may even question in the 3rd person, i.e. not asking participants what they do or think, and as such it may not be critical to the analysis, though there can be much revealed at this point. The discussion moves onto one or two transition questions, which move the discussion towards the key questions and issues. Transition questions help participants to envision the topic in a broader scope. Through these questions, participants become more aware of their fellow participants. Transition questions can ask participants to go into more depth about their experiences and uses of a product, making the connection between the participant and the topic under investigation[xv]. The moderator can then move onto the key questions developing specific questions, issues and probes that can advance as the moderator tunes into the dynamics of the group. There may be additional, new issues that develop and, indeed, issues that group members do not see as being appropriate, and these can be discussed. The emphasis should be upon an evolution and learning process rather than administering a consistent set of questions.
Topic guide
A list of topics, questions and probes that are used by a moderator to help them manage a focus group discussion.
The types of group members to take part in the discussions are then specified. From this specification, a questionnaire to screen potential participants is prepared. Typical information obtained from the questionnaire includes product familiarity and knowledge, usage behaviour, attitudes towards and participation in focus groups, and standard demographic characteristics. With the types of participants specified, consideration must be taken of what would make them relaxed and comfortable, balanced both in a physical and psychological sense.
Having decided on the location of the focus groups, the actual recruitment of group members progresses. This is one of the most difficult tasks, as potential group members may be skeptical of what may happen at the group, sometimes fearing that they are exposing themselves to a hard-sell campaign of time-share holidays or home improvements! If individuals have attended a focus group beforehand, the process of recruitment is easier, but getting the right type of participant together at the right place and time can prove difficult. With the screening questionnaire, recruitment may take place on a face-to-face basis through street interviews or through database details by telephone. One traditional approach is to give the specification of group members to an individual in whose home the discussions are to take place. That individual then recruits participants who fit that specification from their local community. The advantage of this approach is their ability to persuade participants that the process is a bona fide research process and is going to be rewarding in many ways; ultimately they make sure that potential participants actually attend. The big disadvantage is ensuring that those recruited match the screening questionnaire requirements. Whichever method of recruiting participants is used, even when individuals have said that they will participate in the group, a telephone follow-up is necessary to remind and motivate group members.
Group participants have to be rewarded for their attendance. Usually they enjoy the experience immensely once they are there, but that does not ensure that they attend in the first place. Attendance can be rewarded with cash, a donation to charity or a gift. The following example illustrates the difficulties involved in recruitment and a marketing researcher’s creative solution to the problem.
example
So how do you upstage a Ferrari owner?[xvi]
Researching an upmarket, socially active audience is difficult at the best of times. The target is opinionated, demanding, often resistant to research and almost impossible to reach. So, when we got the brief to conduct focus groups among Ferrari, Porsche, top Mercedes and other exotic sports car owners, we were tempted to panic. We knew we could find them, but how could we persuade them to participate?
We realised the one thing that would link our target, who were also defined as keen drivers, not just poseurs, was their love of cars and desire to know the latest news about new models (and to try them out if possible).
That’s why we decided to offer the carrot of a drive around a race and proving track and the opportunity to meet the design and management team at our famous sports car maker. If anything might motivate people, who clearly already had sufficient money to indulge a very expensive taste, it should be this package. It worked like a dream, and we had great success getting the right people to come and, more importantly, to participate.
The first focus group to be run should be seen as an experimental group. All aspects of running the group should be evaluated. Were the group members relaxed and comfortable in the chosen location, i.e. did the context work as intended? How did they react to the tape recorder, video or two-way mirror, i.e. at what point did they seem to relax and forget that they were being recorded? What types of member interacted well or not, and what issues helped or hindered interaction? How did the introductory and transition questions work in opening up and developing the discussion? How did the topic guide work, were there issues missing or issues that individuals would not tackle? How comfortable was the moderator handling the topics, did they have to interject to liven the discussion? How much did the moderator have to know about the subject to have credibility with the participants? With a reflection of these issues, any necessary alterations can be made to the way that the remaining focus groups are administered. There may be very useful information that emerges from the experimental group that can be included in the main analysis. However, if the group does not work well, the information gleaned may be of little use but the lessons learnt are invaluable in running the remaining groups.
Experimental group
An initial focus group run to test the setting of the interview, the opening question, the topic guide and the mix of participants that make up a group.
Finally, the focus groups can be actually run. The question arises here of how many groups should be run. Beyond the first experimental group, the number of groups needed can vary. The extent to which comparisons are sought in analyses can determine how many groups are needed. Seeking comparisons means recruiting participants with different backgrounds or experiences. If there are a great variety of types of individual that make up a target market, then many homogeneous groups may be needed to reflect the variety of types, e.g. a group of 18–25-year-old single male car owners compared with groups of women or groups of older males, married men or non-car owners. The definition of these distinct target groups to question is entirely bound in the nature of the research problem.
A confounding factor in the definition of these groups is the extent to which different types of participants will mix together in a single group. In an experimental focus group that explored attitudes and behaviour related to sports activities in the English city of Bath, distinct target groups of participants did not work well together. Older participants who participated in more ‘gentle’ sports activities did not particularly appreciate, listen to or respect the views of younger participants. In addition, there was a gender split in that male participants showed very little respect for the views of female participants. The older male participants were particularly patronising to younger females. As a result it was decided to run groups that separated younger and older participants and males and females. This meant an increase in the total number of focus groups conducted but also meant that the remaining focus groups worked well and that the views of distinct target groups were clearly presented.
If target participants are geographically widespread, then many groups may be needed to represent this diversity. The following example of an AC Neilsen hypermarket study illustrates the comparisons made in focus group analyses. The important markets they concentrated upon were Russia, Hungary, Poland and the Czech Republic and so focus groups would be needed to represent each country. Further analyses of two distinct target groups were also needed. Thus, a minimum of two groups per country, i.e. eight groups plus an experimental group, would be needed, which would be doubled if it were subsequently felt to be important to run exclusively male and female groups.
example
Hypermarkets in Eastern Europe and Russia[xvii]
International retailers have pounced upon the opportunity to dominate grocery retailing in Central Eastern Europe and Russia. Given the inefficiencies of many local networks, this has created enormous changes in the way that people shop. Never before have they had such choice in store numbers, formats and product assortment. To better understand these changes, AC Neilsen conducted eight focus group discussions with housewives aged 25-40 years in Moscow, Warsaw, Prague and Budapest. Participants who were active shoppers visiting at least three retail formats every month, provided further insights through self-completion questionnaires.
Hungarians were the most enthusiastic about hypermarkets, which they might visit up to twice a week, compared to the fortnightly trip for Czechs. Russians are more budget minded and perfer no-frills discounters. Shoppers in Poland were the most satisfied with hypermarkets although this could be because they were a relatively new phenomenon.

Another factor to be considered in determining the number of focus groups to conduct relates to whether the researcher is using focus groups as part of a grounded theory approach. With such an approach (as described in Chapter 6) theoretical sampling is adopted whereby further examples or instances are sought that may contradict the nature of the emerging grounded theory. The complexity of the developing grounded theory and the extent to which the qualitative researcher makes sense of the issues they are exploring will ultimately determine the number of discussions needed. This can be contrasted to focus groups conducted from a positivist perspective, focusing upon generating an understanding of issues to be confirmed in a subsequent survey. In the latter scenario, the researcher may not need to continue to search for more contradictory perspectives. Whichever paradigm underpins the application of focus groups, resources permitting one should conduct additional discussion groups until the moderator can anticipate what will be said. The last sentence is a reminder of the final factor determining the number of discussions – the time and money afforded by the client.
In summary, the number of focus groups that should be conducted on a single subject depends on the following factors.
§ The extent to which comparisons are sought
§ The different types of participant to be targeted and how well they mix together
§ The geographic spread of participants
§ The paradigm that underpins the focus group
§ The time and budget available.
Another dimension of running the groups, beyond the actual number of groups, is the nature of stimuli that the moderator chooses to input. At certain times, examples of particular products or brands may be introduced for participants to examine, or even taste or sample if the nature of the product permits. Advertising material such as brochures, posters or even video recordings of television or cinema adverts can be shown and a response generated. One of the most frequently used forms of stimuli is the mood board.
Mood board
A collage created in a focus group setting. Focus group participants are asked to snip words and pictures from magazines that they see as representing the values a particular brand is perceived to have. In some circumstances, collages can also be made up from audio and video tapes.
The ‘mood board’ is really the creation of a collage. Focus group participants are given a pile of magazines and asked to snip words and pictures from them. The direction they are given is to select images and words that they think represent characteristics of a brand, a consumer type or lifestyle or whatever issue the researcher wishes to be illustrated in this manner. The resultant collage can then be used to stimulate discussion and to help draw together ideas and connect them. Creative marketing decision-makers in copywriting or advertisement development may develop many ideas directly from the collages or mood boards. The mood board has two main functions:
§ Reference point. The moderator can use the mood board to reflect upon the discussion, in which case issues can emerge that were not so apparent in the heat of a discussion.
§ Enabling device. The mood board gets participants to loosen up and talk more freely. The focus group is not to get participants to talk rationally but to display what ‘feels right’ to them. The collage can help to express feelings they may not be able to put into words, or enable those words to have more clarity. This can happen by questioning what is included in a mood board as well as what is omitted. [xviii]
The following example illustrates that the mood board can develop beyond two-dimensional images.
example
Art of the matter[xix]
Paul Walton, Chairman of new product development consultancy The Value Engineers, explains: ‘In the early stages of a new product or brand reassessment project, words might be appropriate. Beyond the words come picture collages and as you learn more and you start to give the brand a clearer identity, you introduce mock-ups of packaging and other props to ‘three-dimensionalise’ the world.’
Alex Authers, Research Director at branding consultancy New Solutions, argues the case for stimuli beyond the mood board: ‘We’ve moved on from static visuals, to using videos. It’s often useful to show a series of fast-edited clips set to some sort of soundtrack. People are now much more video literate. So, instead of having a mood board, we have a mood video. Video collages are particularly good for exploring the emotional resonances of brands.’
Other props used in qualitative research include swatches of material and even fragrances. Authers says smells and colours can ‘help to create a mood and evoke a positioning’.
The final stage in planning and conducting focus groups involves the analysis of data. Chapter 9 discusses qualitative data analysis in more detail. However, at this point there are two essential points to note:
1. Evolving analysis. Focus groups can change and develop in terms of the issues discussed and the stimuli used to draw out views from participants. The changes are made as the moderator generates and develops new ideas as each focus group progresses. The moderator makes observations and notes to help them as the discussion progresses and also for when they are over. These observations and notes are part of the total analysis in that they decide which issues to probe, which issues to drop and the form of summarising issues that may be presented to groups at certain stages of the discussion.
2. Not just the narrative. If the discussion is recorded then transcripts can be produced which can be analysed with proprietary software. These transcripts form a major part of the analysis procedure but the accumulation and reflection upon observations and notes forms a key part of the analysis.
The moderator
Throughout this chapter we have referred to the moderator as an individual who conducts a focus group discussion, by setting the purpose of the discussion, questioning, probing and handling the process of discussion. This individual may be the researcher handling the project. More likely it will be someone who specialises in the technique, or given the number of groups to run and the time allowed to complete them, a number of specialist moderators will be employed. Whoever is to undertake the task of ‘moderating’ will require the following qualities:[xx]
1. Kindness with firmness. The moderator must quickly develop an empathy with group members. From this the moderator should show kindness to make participants feel welcome, combined with a firmness to stop particular individuals taking over the discussion.
2. Permissiveness. The moderator must be permissive, allowing the flow of discussion to develop as the group sees fit. However, they must be alert to signs that the group’s cordiality or purpose is disintegrating.
3. Involvement. The moderator must encourage and stimulate intense personal involvement. In certain circumstances, this may mean becoming involved in the actual discussion itself. This can happen if a tendency for ‘group speak’ emerges. ‘Group speak’ happens when little debate or creativity in ideas develops, as particular individuals may not wish to be seen as going against a perceived group norm.
4. Incomplete understanding. The moderator must encourage participants to be more specific about generalised comments by exhibiting a feigned naivety or incomplete understanding.
5. Encouragement. The moderator must encourage unresponsive members to participate.
6. Flexibility. The moderator must be able to improvise and alter the planned outline amid the distractions of the group process.
7. Sensitivity. The moderator must be sensitive enough to guide the group discussion at an intellectual as well as emotional level. They must also be attuned to mood changes and issues that fire up enthusiastic responses or conversely cause the discussion to dry up.
8. Observation. As the group progresses, notes must be made of ideas or questions to come back to, interpretations of particular silences or bouts of laughter, and how group members are interacting with each other. These observations help the group discussion to progress well and the interpretation of the discussion to have greater meaning.
Other variations of focus groups
Focus groups can use several variations of the standard procedure. These include:
§ Two-way focus group. This allows one target group to listen to and learn from a related group. In one application, physicians viewed a focus group of arthritis patients discussing the treatment they desired. A focus group of these physicians was then held to determine their reactions.[xxi]
§ Dual-moderator group. This is a focus group discussion conducted by two moderators. One moderator is responsible for the smooth flow of the session, and the other ensures that specific issues are discussed.
§ Duelling-moderator group. Here also there are two moderators, but they deliberately take opposite positions on the issues to be discussed. This allows the researcher to explore both sides of controversial issues. It also encourages participants who may support a particular perspective to express their views without the fear that they will be ‘attacked’ by the rest of the group.
§ Participant-moderator group. In this type of focus group, the moderator asks selected participants to play the role of moderator temporarily to improve group dynamics.
§ Client-participant group. Client personnel are identified and made part of the discussion group. Their primary role is to offer clarifications that will make the group process more effective.
§ Mini group. These groups consist of a moderator and only four or five participants. They are used when the issues of interest require more extensive probing than is possible in the standard group of 6 to 10.
§ Telephone focus groups. These are conducted using a telephone conferencing system with the same number of participants as conventional focus groups, but typically within a narrower time frame, no more than an hour. These can work well in gaining access to widely dispersed experts in a range of professions or specialists. They would not be cost effective with ‘average’ consumers, except perhaps in cases where follow-up from survey respondents is desired[xxii].
Other types of qualitative group discussions
Brainstorming
Traditional brainstorming[xxiii] has been used for several decades, especially in the context of management or marketing issues. Whether formal or informal, the process is the same: think of as many ideas as you can and say them out loud; leave the evaluation until later; build on and combine others’ ideas; be as imaginative as possible, the wilder the ideas the better. The group moderator seeks to nurture an atmosphere of creativity, tapping into the intuition of participants, generating novel ideas and connections between ideas.
When it works well, ideas flow freely from an interplay that may never have occurred if the group had not brainstormed together.
Two problems plague traditional brainstorming: production blocking and evaluation apprehension.
§ Production blocking occurs when a group member has an idea, but someone else is talking. When it’s their turn, they have forgotten the idea, or think their idea is redundant or not that good. If the group is large or dominated by talkative people, they lose interest and do not say what they think.
§ Evaluation apprehension occurs when participants become anxious about what others think of their thoughts. Ideas may be censored, as there is a fear of being labeled as odd. When participants feel this apprehension, they do not produce as many new and potentially useful ideas but keep them to themselves and therefore defeat the purpose of brainstorming.
Industrial group discussions
As noted earlier, the focus group has limited use in industrial or business research. Getting managers together at the same time and place is a big operational problem. Getting them to be open about their companies in a group scenario is also very difficult to achieve. The following example illustrates how this problem was overcome, essentially by not using the term ‘focus group’.
example
Is a ‘workshop’ a focus group?
In a major pan- European business-to-business survey, group discussions were seen as important in order to explore and develop issues that could not be measured in a questionnaire. What was measured in the questionnaire produced many statistics whose analyses necessitated further exploration and elaboration. Developing an elaboration of the statistical findings could be tackled in a group discussion. To overcome the above issues of getting managers to sit down together, part of the incentive to complete the questionnaire was an invitation to take part in a ‘closed forum for questionnaire participants’. This meant that a date was set to present findings from the survey exclusively to questionnaire participants. The date and meeting place of The Management Centre in Brussels were established well in advance to allow managers the chance to put the date in their diaries. The day started with an initial session of presenting statistical findings to the whole group of managers who had responded to the questionnaire. As questions were fielded in this forum, particular topics of interest were identified and focus groups named as ‘workshops’ were built around these topics. By mid-morning, groups of around 10 managers were together tackling a theme of importance to them. With a loose set of topics to develop into questions and probes, the format for a focus group was achieved. In the afternoon, the same format continued with a presentation of questionnaire findings and further group discussions. By splitting the day in this manner, managers could attend and contribute to two subject areas of interest to them. Finally, individual managers could be identified who could be interviewed in depth on their own at a later date’

Misconceptions about focus groups
There has been much debate about the use and value of focus group discussions. Much of the debate has been mis-informed through the development of myths and misconceptions built around the technique. Misconceptions about focus groups have emerged from assumptions made by researchers or the users of focus group findings or even journalists. Their assumptions may have been useful at one time, in a particular context, or when tackling a certain subject. The assumptions may well apply to specific contexts but may not be applicable to focus groups in general. Working through the misconceptions can help marketing researchers make decisions about when and how to use focus groups, and even if they should be using them at all[xxiv]. The misconceptions can be summarised by the five Es.
1. Economical - they are low-cost and quick. The idea that focus groups are low cost and quick has emerged through comparison with other research methods. In comparison with survey research, focus groups may be conducted with a smaller budget and more quickly. This is not a fair comparison given the sample sizes the techniques work with and the type of data they generate. In comparison with other qualitative methods, focus groups are often more expensive than observation and individual interviewing, primarily due to more expensive recruiting methods. The biggest resource involved in the technique however is that of the researcher in the careful planning, execution and analysis of focus groups. The analysis of the data generated can be ‘cheap and quick’ if it is a case of decision-makers taking away their own conclusions from focus groups they have observed. Analysis can be far more expensive and time consuming if the careful planning and execution is followed up by the assembly of data, reflective diaries and memos and the full immersion into that data of the researcher as detailed in Chapter 9.
2. Experts - they require professional moderators. The qualities of a good moderator were detailed above. A professional moderator with experience of tackling different subjects, questioning and probing different types of participant and working in a variety of contexts, undoubtedly has much to offer any project. However, in many instances it is not the amount of experience that matters most in moderating. Sometimes a less experienced moderator who has more contact with the issues under question, more contact with the participants, and is perhaps more comfortable in the interview location, can elicit better data. This is especially true when working with distinctive ethnic, linguistic or cultural groups. For example, an undergraduate student working on a project for a leading deodorant brand was about to enter the room to conduct a focus group. Looking at the 18-20 year olds before he entered the room, he realised that he was wearing the ‘wrong’ logo. With a quick change of sweatshirt, he continued with the discussion to great effect. Upon reflection, he revealed that the logo he was wearing and indeed the colour of his sweatshirt would have said things about him to the group that would have affected his credibility. He felt that he would not been able to engage with the group and got them to reveal so much in his original clothing. His acute sensitivity to this cultural nuance made this focus group particularly effective, an older and more experienced moderator may not have picked this up. The best moderators are not necessarily those with the most experience. It is the individual that can best engage with, listen to, and draw out the best of target participants.
3. Easily upset - they don’t work for sensitive topics. Focus groups are regularly used in projects related to sexual behaviour, substance abuse and stressful life events, which have marketing or other social implications. This misconception is based upon what may be deemed as socially acceptable in conversations, perhaps around the dinner table. However, focus groups present an atypical setting for discussion. The moderator is encouraging everyone in the group to share values and experiences that they all have an interest in. There may be little consequence to what they say, especially if they are meeting with strangers that they may never meet again. Handled with caution, the moderator can encourage participants to reveal things they would normally keep to themselves. Researchers and moderators working with sensitive issues must make plans to encourage appropriate self-disclosures, and to cope with disclosures that go beyond the boundaries of the project. These plans go beyond the wording of the questions and probes, into the atmosphere they wish to create at the discussion, the nature of the location being particularly important here.
4. Endorsement - they must be validated by other research methods. This is part of the general misconception that all qualitative techniques play a preliminary role that prepares the way for a conclusive technique. Focus groups can serve a useful role as the first stage of developing questionnaires or experiments, but that is not their sole purpose. Focus groups can illuminate and indeed help to validate many of the statistical findings that emerge from surveys, for example helping to understand and describe findings from factor and cluster analyses. They can be used in isolation to produce the kind of in-depth understanding of an issue that no other technique can provide. For example, if the use of a particular celebrity in an advertisement can generate a series of emotional responses that could impact upon the brand values – and those values can compounded by the particular use of music and who is playing that music. To uncover what is happening to the viewers and how their values may have changed, the focus group may well be the most efficient and effective technique, of all qualitative or quantitative options.
5. Exposure - they reveal how consumers will behave. Focus groups, depth interviews and surveys all depend upon verbal reporting. They depend upon a belief in what participants say about how they intend to behave. In many circumstances, even when it comes down to it, the most sincere participants can change their minds. In focus groups, participants may talk about their likely behaviour in many hypothetical scenarios, moderators can watch other participants nod their heads in agreement and the evidence is compelling. It must be recognised that the data generated is attitudinal, and trying to predict behavior from attitudes is most problematic. The rise in customer databases and ethnographic methods has helped marketing researchers to predict consumer behaviour and validate the findings of focus groups and surveys.

The focus group, like every research technique is never foolproof or perfect. To get the most from them, the following thinking should be encouraged[xxv]. First, the quality of moderating is crucial to the success of the discussion. The quality of the findings is directly related to the talents, preparation and attentiveness of the moderator. Second, teamwork is vital. As well as star moderators, there must be quality recruiters, note takers, analysts and reporters. Third, this team has always something to learn from the participants. For a short period of time, the participants open their lives (and sometimes souls) and share their experiences, preferences and beliefs. The most destructive thing a researcher can do in a focus group is to display arrogance, condescension or superiority. It is much better to sit back and let participants tell you everything you wanted to know, and more! Finally, there are many ways to conduct successful focus groups. Given the nature of the topic under investigation and the nature of the target participants, there are many options to choose of the best context and approach to draw out the best from participants. Focus groups vary enormously in how they are planned, administered and analysed, from the very structured interview in a studio bristling with technology, to a passionate dialogue conducted globally over the Internet, to a riotous exchange on a tropical beach. Experimentation with new ways of conducting focus groups should be positively encouraged rather than being bound by the thought that there is one ideal way to conduct them.

International marketing research
The term ‘focus group’ is commonly used across all continents, yet it subsumes approaches that are different. For many European marketing researchers, one of the biggest headaches associated with focus groups is its name. The technique has been disparaged by a popular media impression that implies that focus groups are a means to get answers to set questions. Many see this as the US approach to conducting focus groups and thus see the name as an unwanted Americanism that has displaced what they see as the favoured term, ‘group discussion’.
The best people at moderating focus groups are ones who can create that spirit of spontaneity. You can’t do it with a crowded agenda. You can’t do it if you’re focused, which is why focus is wrong.[xxvi]
The above quotation implies a criticism of a US approach that can be highly structured rather than an approach that is truly spontaneous and exploratory in its fullest sense. It is not advocated at this stage that the word ‘focus’ be removed, but it is a reminder that there are different research approaches that can affect how the technique is administered. US and European approaches were presented in Chapter 6 to illustrate how different philosophies that underpin research techniques can affect how the technique is applied.[xxvii] In examining the use of focus groups, two main schools of thought were presented: the cognitive approach, which largely follows a format and interviewing style as used in quantitative studies and is generally used by American researchers; and the conative approach, which has less structure to the questions, with group members being encouraged to take their own paths of discussion, make their own connections and let the whole process evolve, and is generally used by European researchers. It is not advocated that one particular approach is better than the other. Each approach has its own distinctive strengths and weaknesses depending upon why the technique is being used, the nature of the participants and the nature of researchers and decision-makers who are to use the data.
For example, in the USA, observing focus groups through the use of purpose-built focus group rooms or a viewing laboratory with a two-way mirror is a policy. Typically there will be five to eight observers, sometimes more, with agency and marketing company matching level for level to maintain a balance of power. American researchers and clients defend the value of observers. People with different perspectives can listen in a way that a moderator cannot, since the moderator is often ‘dipping in for the moment’ on one specific issue while the clients ‘have the brand in their bones’. Probably the most deeply felt reason for being there is just being there; there is no substitute for the touchy-feely benefit of experiencing the consumer first-hand. More than ever, marketing people are isolated in their small worlds (probably in atypical New York or Chicago), making assumptions about their users. Seeing them is a reality check. Hearing tonality, watching body language, observing the consumer interact with the product, enriches the learning process. Sometimes brand people will get rejuvenated and creatives will get inspired.[xxviii]
Viewing laboratory
A room where a focus group may be conducted and simultaneously observed, usually by using a two-way mirror.
example
A client’s view of observing the focus group through the two-way mirror[xxix]
‘It’s a strange experience, sitting on the other side of a two-way mirror watching your product being analysed in a focus group. It’s like being locked out of your house and peering through the windows as burglars rifle through your drawers and take the piss out of your record collection.’
‘You want to bang on the glass and plead with them to stop, but then they say something nice about your product and you want to grab the person next to you and shout “I thought of that!”.’
‘You tell yourself not to take the criticism too personally, to be professional, but it’s hard when you’ve had sleepless nights agonising over minute details only to see them ignored or ripped apart in the space of half an hour.’
‘It’s great to get raw feedback like this, but remember just how raw it is. One group can say completely different things to another and some may just like the sport of making the suits behind the mirror suffer.’
At face value, the above description of the use of viewing labs seems to show a feature of US practice that has many benefits, but consider the following problems:
§ Incomplete data. Clients may not have seen all the groups, leaving gaps in the total picture. Observers admit that they focus on the most lively and self-serving points; that they identify positively and negatively with certain responses and filter accordingly; that they are attentive to ideas that affirm their positions and dismissive of contrary viewpoints.
§ Instant analysis. Rather than waiting for a moderator’s report or even the conclusions of the group, observers often jump to conclusions on little evidence or, worse still, stop listening once they have an ‘impression’ of the results.
§ Moderator reflection. Often the moderator gets trapped into debriefing immediately after the group and taking positions that are hard to retreat from but which might be very different after a thoughtful review.
§ Effect on participants. There is little research on the effect of the two-way mirrors on participants. The setting is not exactly conducive to natural expression and participants may alter their responses for effect, stay silent or be self-conscious.
Supporters of the European approach to focus groups may console themselves that they can overcome these drawbacks. They may argue that the context in which they run groups is more conducive to relaxed participants, that moderators are not under such pressure to produce ‘instant analysis’ and that they can take more time to reflect and interpret the data generated. But in Europe, and especially in Britain (which has had the biggest tradition of running focus groups in people’s homes), marketing researchers are embracing the use of the viewing lab. Marketing researchers should not take a dogmatic position, believing that their approach is the most ‘correct’. The marketing researcher should be open-minded to learn from focus group practices in different countries and be able to critically evaluate why they have been administered in that way.
Ethics in marketing research
A number of ethical issues related to running focus groups have emerged in this chapter. The focus group can be a direct qualitative research technique where the purpose of the discussion is made clear to participants before it starts. However, the focus group can incorporate observational and projective techniques that introduce elements of indirectness, i.e. elements or all of the purpose of the discussion being hidden from participants. This is where the marketing researcher faces a real ethical dilemma. If they fully reveal the purpose of their study, would this put off potential participants whose views are important to the success of the study? Even if they do manage to recruit potential participants, the researcher has to consider how they may feel when the real purpose of the study becomes apparent through the nature of the discussion or by the moderator revealing all at the end. The nature of the dilemma faced by marketing researchers is that by revealing too much at the start of the study they may compromise the quality of their discussion. A full revelation may not be conducive to participants reflecting, making new connections and expressing themselves about issues that may be deeply held, difficult to conceive and express.
Researchers should take all reasonable precautions to ensure that participants are in no way adversely affected or embarrassed as a result of the focus group. What is meant by an adverse effect and embarrassment will largely depend upon the issues being explored and how they are perceived by target participants. Some participants may find personal hygiene issues very embarrassing to talk about in a group scenario, but not financial issues. Other individuals may be very frank about their sexual behaviour, while others would be shocked at the notion of talking about sex with a group of strangers. The researcher must get to know how the issues they wish to explore are perceived by their target participants by examining literature, secondary data and the use of experimental focus groups.
Another major ethical problem that is research issue and participant specific is the use of alcoholic drinks during focus groups. For certain groups of participants, relaxing and socialising in a comfortable context, drinking wine or beer is very natural. Researchers with experience of running many focus groups would argue that serving alcoholic drinks can help to reduce tension in certain participants and give them the confidence to express their particular viewpoint. Other researchers would argue that this is unethical practice, that in effect the researcher is ‘drugging’ the participants. Whatever the researcher decides is right for the type of participants, the issues they are questioning them about and the context in which the discussion takes place, there are practical problems involved with serving alcohol. Controlling the flow of alcohol and how much is given to certain participants may take attention away from the discussion. If control is not exerted, particular participants may get out of hand and disrupt or even destroy the discussion. Researchers could be accused of not taking reasonable precautions to ensure that participants are in no way adversely affected or embarrassed as a result of the focus group, should the use of alcohol be abused by certain participants.
Finally, in no circumstances should researchers use audio or video recording or two-way mirrors in focus groups without gaining the consent of participants. It must be made clear to participants that recording or observation equipment is to be used and why it needs to be used. Participants must then be free to decline any offer to take part in a discussion.
Internet and computer applications
The development of the Internet, the increasing numbers of individuals with access and who are comfortable using the Internet, has presented many opportunities to run focus groups online. For online focus groups, virtual facilities can be used, providing the same facilities as real-life facilities including ‘rooms’ such as a reception room, discussion room and client backroom. Online focus groups can be conducted in two ways: ‘real-time’ or ‘non-real-time’.[xxx]
§ Real-time. In this method, all participants are online at the same time. The transmission of messages is immediate or as close to immediate as can be. As one participant types in their message, it is transmitted to the group as a whole. Other participants can read the message and can reply as soon as they receive it. Real-time focus groups can be highly interactive with a fast ‘passionate’ exchange of views. This is the big advantage of real-time groups, that the passion for a subject and the nature of rapid exchange can develop a very creative atmosphere. Participants do not have to wait for others to comment in order to send further messages, so as they have been stimulated to think in particular ways, their thoughts and ideas can flow straight out on to the screen. The drawback is that the participant with the most proficiency at typing has the power to dominate the discussion. Compared with face-to-face focus groups where the moderator and audio and video recordings can track participant statements and the responses stimulated, this cannot be done online in real time. The distinction of replying and sending becomes a blur as participants do not take their turn as they would face to face. This means that much of the structure of conversation is lost.
§ Non-real-time. In this method there is no requirement for participants to be online at the same time. This method is analogous to the use of email. Unlike email, though, this form of focus group is conducted using a ‘conference site’ as opposed to individual email addresses. Participants can send their views about a particular issue to a conference folder. Messages can be read in real time but this does not usually happen. The reality is that all responses are archived in the folder and can be opened and responded to by other participants. This method of running focus groups can overcome difficulties in running focus groups in different time zones and therefore is an excellent tool in conducting international research. Participants with weak typing skills do not lose their voice or feel intimidated. The major benefit of this approach is that akin to face-to-face focus groups, participants can ‘sit back’ and reflect upon how they really feel about an issue. Even better than in face-to-face scenarios, they can take time to express how they really feel about an issue.
Online focus groups have a number of logistical requirements. Depending upon how the online focus group is to be conducted, participants may have to gain access to proprietary conferencing software. Real-time and non-real-time focus groups can operate from particular Websites, requiring potential participants to have a Web browser and to know the Web address from which the group will be run.
In summary, the key benefits of running focus groups online include the following:[xxxi]
§ More potential participants can be recruited through the growing use of the Internet, and the growing ease of conducting discussions online.
§ Participants can be made to feel that they have the ability to contribute; their confidence can be quickly built up.
§ Conflicts in face-to-face focus groups that may stem from participants taking a dislike to other participants from their physical appearance can be avoided.
§ A great breadth of information may be collected, through the types of participant that can be recruited and the geographic spread of participants.
§ The practical difficulties of getting individuals together at the same time in the same location can be overcome.
§ The nature of a discussion location that is ‘comfortable’ to the participant is largely overcome by each participant setting the conditions that they feel comfortable in.
Summary
In direct qualitative methods, participants are able to discern the true purpose of the research, whereas indirect methods disguise the purpose of the research to some extent. Focus groups can be a completely direct qualitative method, though by the use of observational and projective techniques elements of indirectness can be introduced. In deciding how ‘direct’ a focus group should be, researchers have to face ethical issues and questions related to the richness of data that they can draw from participants. Focus group discussions are the most widely used qualitative research technique.
Focus groups are conducted by a moderator in a relaxed and informal manner with a small group of participants. The moderator leads and develops the discussion. In a focus group, participants can portray their feelings and behaviour, using their own language and logic. The value of the technique lies in the unexpected findings that emerge when participants are allowed to say what they really feel. An experimental focus group should always be run at the start of any series of discussions. The researcher needs to understand how comfortable target participants feel in the chosen location for discussion, how the opening question works, the topic guide, the probes, the mix of participants and the refreshments, including any alcoholic drinks. In a nutshell, one could argue that one should continue running focus groups until nothing new is learned from target participants. This perspective oversimplifies how diverse and spread out different participants may be and how well they mix together in a group situation. The time and budget available will also mean that a cut-off point has to be drawn and the strongest analysis and interpretation has to be made at that point. Development in the use of the Internet means that group discussions can take place away from the traditional ‘viewing’ or meeting room. The maxim of participants being relaxed and comfortable with the context and means of communicating still applies.
There are a variety of different styles of running focus groups. The variations and adaptations are used to draw the best out of particular types of participant, tackling particular types of issue. The most significant development in the variation of technique has been afforded by developments in Internet technology and take-up. Participants can be targeted from all over the world, discussing issues online in ‘real-time’ or ‘non-real-time’. Another factor that affects the style of running focus groups lies in the underlying philosophy for conducting research. US and European styles can be broadly encapsulated as being respectively ‘highly structured’ or ‘highly spontaneous’. There is no one absolute correct method to administer a focus group; the researcher should understand the factors that will make the technique work for their particular research problem and the type of participants they have to work with. A key element of ensuring that the focus group works well lies in the ethical issues of how much is revealed to participants before they get involved and during the discussion. Getting participants to relax and be open may involve the use of alcoholic drinks that for many researchers creates no problems but for some creates ethical and practical problems.
Questions
1. Why may marketing researchers not wish to fully reveal the purpose of a focus group discussion with participants before it starts?
2. What are the key benefits and drawbacks of conducting focus group discussions?
3. What are the difficulties in conducting focus groups with managers or professionals?
4. What determines the questions, issues and probes used in a focus group?
5. Evaluate the purpose of running an experimental focus group discussion.
6. What does a ‘comfortable setting’ mean in the context of running a focus group?
7. To what extent can a moderator achieve an ‘objective detachment’ from a focus group discussion?
8. Why is the focus group moderator so important to the success of a focus group discussion?
9. What are the relative advantages and disadvantages of being able to covertly observe a focus group discussion?
10. What can the researcher do to make potential participants want to take part in a focus group?
11. What determines the number of focus groups that should be undertaken in any research project?
12. Describe the purpose and benefits of using stimulus material in a focus group.
13. What is the difference between a dual moderator and a duelling moderator group?
14. Describe the opportunities and difficulties that may occur if alcoholic drinks are served during focus group discussions.
15. Evaluate the benefits and limitations of conducting focus group discussions on the Internet.
Exercises
1. Following the methods outlined in this chapter, develop a plan for conducting a focus group study to determine consumers’ attitudes towards organic foods. Specify the objectives for the groups, write a screening questionnaire, list potential props or physical stimuli that you could use in the discussion and develop a moderator’s outline.
2. The campus sports centre is trying to recruit more members from the local non-student community. In achieving this aim, evaluate the marketing decisions that could be supported by focus groups, either as a technique in its own right or validated with other techniques.
3. You are a brand manager for Johnny Walker whiskies. You wish to invest in an Internet focus group study of whisky buyers and drinkers. Explain how you would identify and recruit such respondents from across the globe. What incentive(s) would you offer potential participants?
4. Visit the website of the Association of Qualitative Research Practitioners (www.aqrp.co.uk). Examine the reports and views of contributing practitioners and write a report on what you feel are the latest developments and/or exciting opportunities in the use of focus groups.
5. In a small group discuss the following issues: “the dress, appearance and speech of the moderator create biases in group discussions that cannot be evaluated” and “mood boards created in focus groups are more useful to marketing decision-makers compared to a formal written analysis of the discussions”

Figure 7.1 A classification of qualitative research techniques
Clarify marketing research problem(s) and objectives
¯
Clarify the role of focus groups in fulfilling those objectives
¯
Specify the issues to be developed in the focus groups
¯
Specify the types of target participants to make up groups
¯
Specify the location(s) in which to conduct the focus groups
¯
Recruit group members
¯
Run an experimental group
¯
Conduct the focus groups
¯
Analyse data and present findings
Figure 7.2 Procedure for planning and conducting focus groups
Notes

[i] Murphy, D., ‘What can be more real than a result?’ Research World, (December 2004), 19
[ii] Morgan, D.L. ‘Focus Group Interviewing’, in Gulbrium, J.F. and Holstein, J.A., (eds.) Handbook of Interview Research: Context and Method, (Thousand Oaks, CA: Sage 2002) 141-159
[iii] Morgan, D.L. and Krueger, R.A. ‘When to use focus groups and why’, in Morgan, D.L, (ed.) Successful focus groups: Advancing the state of the art, (Newbury Park, CA: Sage 1993) 3-19
[iv] ESOMAR, Industry Study on 2004, Esomar World Research Report, 2005, p. 26
[v] Bloor, M., Frankland, J., Thomas, M. and Robson, K., Focus groups in Social Research (Thousand Oaks, CA: Sage Publications, 2001).
[vi] Kreuger, R.A., and Casey, M.A., Focus Groups: A practical guide for applied research, 3rd ed. (Thousand Oaks, CA: Sage Publications, 2000). Drayton, J. and Tynan, C., ‘Conducting focus groups – a guide for first-time users’, Marketing Intelligence and Planning 6(1) (1988), 5–9.
[vii] For more discussion, see Morgan, D.L. ‘Planning focus groups’, (Thousand Oaks, CA: Sage 1998) 71-76; Dachler, H.P., ‘Qualitative methods in organization research’, Organizational Studies 18(4) (1997), 709–24.
[viii] Forrest, C., ‘Research with a laugh track’, Marketing News, 36, (5) (March 4 2002) 48, Mazella, G.F., ‘Show-and-tell focus groups reveal core bloomer values’, Marketing News 31(12) (9 June 1997), H8.
[ix] Morgan, D.L. ‘Planning focus groups’, (Thousand Oaks, CA: Sage 1998) 71-76
[x] MacDougall, C., ‘Planning and recruiting the sample for focus groups and in-depth interviews’, Qualitative Health Research, 11 (1) (January 2001) 117-126; Kahn, H., ‘A professional opinion’, American Demographics (Tools Supplement) (October 1996), 14–19.
[xi] Gordon, W., ‘New life for group discussions’, ResearchPlus (July 1993), 1.
[xii] O’Donoghue, D., ‘Taking the plunge’, Research, (December 2005) 46-47 .
[xiii] Krueger, R.A., Moderating focus groups, (Thousand Oaks, CA: Sage 1998) 30
[xiv] Goldsmith, R.E., ‘The Focus Group Research Handbook’, The Services Industries Journal, 20 (3) (July 2000) 214-215; Greenbaum, T.L., The Handbook for Focus Group Research (Newbury Park, CA: Sage, 1997).
[xv] Krueger, R.A., Developing questions for focus groups, (Thousand Oaks, CA: Sage 1998) 25
[xvi] Ellis, R., ‘So how do you upstage a Ferrari owner?’, ResearchPlus (November 1994), 10.
[xvii] Ghosh, S., ‘Hypermarkets lead to new shopping habits’, Research World, (November 2004) 32-33
[xviii] Croft, M., ‘Art of the matter’, Marketing Week (9 October 1997), 71
[xix] Ibid.
[xx] Katcher, B.L., ‘Getting answers from the focus group’, Folio: The Magazine for Magazine Management (Special Sourcebook Issue for 1997 Supplement) 25(18) (1997), 222; and an adaptation from Chase, D.A., ‘The intensive group interviewing in marketing’, MRA Viewpoints (1973).
[xxi] Silverstein, M., ‘Two-way focus groups can provide startling information’, Marketing News (4 January 1988), 31.
[xxii] Mariampolski, H., Qualitative marketing Research: A comprehensive guide, (California, Sage 2001) 47
[xxiii] Gallupe, R.B. and Cooper, W.H., ‘Brainstorming electronically’, Sloan Management Review 35(1) (Fall 1993), 27.
[xxiv] Morgan, D.L. ‘The focus group guidebook’, (Thousand Oaks, CA: Sage 1998) 45-52
[xxv] Morgan, D.L. ‘The focus group guidebook’, (Thousand Oaks, CA: Sage 1998) 52-54
[xxvi] Savage, M., ‘Soft focus’, Research (September 1999), 32–3.
[xxvii] Note that the expression ‘US’ or ‘European focus group’ does not mean that such methods are exclusively used in these continents. The term is used to show that there is a greater propensity for a particular approach to running a focus group in a particular manner in each of these continents.
[xxviii] Sonet, T., ‘See the USA, through the looking glass’, ResearchPlus (June 1994), 6.
[xxix] Dwek, R., ‘Through the looking glass’, Marketing (11 September 1997), 37.
[xxx] Mann, C. and Stewart, F., Internet Communication and Qualitative Research: A Handbook for Researching Online (London: Sage, 2000), 101–2.
[xxxi] Sweeney, J.C., Soutar, G.N., Hausknecht, D.R., Dallin, R.F. and Johnson, L.W., ‘Collection of information from groups: a comparison of two methods’, Journal of the Market Research Society 39(2) (April 1997), 397.

Qualitative research: its nature and approaches

2010-06-04T09:03:00.000-07:00

Qualitative research: its nature and approaches
Stage 1 Problem definition
Stage 2 Research approach developed
Stage 3 Research design developed
Stage 4 Fieldwork or data collection
Stage 5 Data preparation and analysis
Stage 6 Report preparation and presentation
Objectives
After reading this chapter, you should be able to:
1. explain the difference between qualitative and quantitative research in terms of the objectives, sampling, data collection and analysis, and outcomes;
2. describe why qualitative research is used in marketing research;
3. understand the basic philosophical stances that underpin qualitative research;
4. understand the nature and application of ethnographic approaches;
5. understand how qualitative researchers develop theory through a grounded theory approach;
6. explain the potential of action research to qualitative marketing researchers;
7. discuss the considerations involved in collecting and analysing qualitative data collected from international markets;
8. understand the ethical issues involved in collecting and analysing qualitative data.
Qualitative research helps the marketer to understand the richness, depth and complexity of consumers.
Overview
Qualitative research forms a major role in supporting marketing decision-making, primarily as an exploratory design but also as a descriptive design. Researchers may undertake qualitative research to help define a research problem, to support quantitative, descriptive or causal research designs or as a design in its own right. Qualitative research is often used to generate hypotheses and identify variables that should be included in quantitative approaches. It may be used after or in conjunction with quantitative approaches where illumination of statistical findings is needed. In some cases qualitative research designs are adopted in isolation, after secondary data sources have been thoroughly evaluated or even in an iterative process with secondary data sources.
In this chapter, we discuss the differences between qualitative and quantitative research and the role of each in marketing research. We present reasons for adopting a qualitative approach to marketing research (Stage 2 of the marketing research process). These reasons are developed by examining the basic philosophical stances that underpin qualitative research. The concept of ethnographic techniques is presented, with illustrations of how such techniques support marketing decision-makers. The concept of grounded theory is presented, illustrating its roots, the steps involved and the dilemmas for researchers in attempting to be objective and sensitive to the expressions of participants. Action research is an approach to conducting research that has been adopted in a wide variety of social and management research settings. Action research is developing in marketing research and offers great potential for consumers, decision-makers and researchers alike. The roots of action research are presented, together with the iterative stages involved and the concept of action research teams. The considerations involved in conducting qualitative research when researching international markets are discussed, especially in contrasting approaches between the US and Europe. Several ethical issues that arise in qualitative research are identified.
Before the chapter moves onto the substantive issues, one key point to note in the application of qualitative research is the name applied to the individuals who take part in interviews and observations. Note the emphasis in the following example:
example
‘Research is War’[i]
Dutch agency MARE is calling for respondents to be promoted to the status of participants in research. The method that they have been developing puts the respondent in the position of participant rather than a passive reactive pawn in research. We take two participants, show them an image or commercial, and then invite one participant to ask the other what he or she just saw. This results in a conversation that is not directed by researchers, which is an important aspect. Our intervention is mostly focused on finding a good match between two participants who can communicate with one another. The method reveals how a consumer absorbs information and reports about it to fellow consumers, and it shows the client which elements of a commercial message work and which elements don’t. Clients were reluctant to use this approach when it was first used in 1995. MARE sense there is a market for it now, so with an amount of refining, adjusting and testing, it will be running in September 2005. A multinational in the Netherlands which has yound marketers will apply it in research among young consumers.■

This example illustrates the creative thinking necessary to get the most from qualitative research and the respect that should be given to individuals who may be asked to engage in a process that sometimes goes way beyond simple questioning. Embracing this attitude, the term ‘participant’ rather than ‘respondent or ‘informant’ is used throughout Chapters 6 to 9, i.e. the core qualitative research chapters of the text.
We now move onto the nature of qualitative research with two examples. Given the nature of its products and competitive environment, the first example illustrates why L’Oréal feel that qualitative research is of importance to them. The second example illustrates how Phillips uses qualitative techniques to support their trend forecasting and product design. Note in this example the use of an analytical framework to help researchers and decision-makers gain insight from the data they collect. These examples illustrate the rich insights into the underlying behaviour of consumers that can be obtained by using qualitative techniques.
example
A research commitment more than skin deep[ii]1
L’Oréal is the largest supplier of toiletries and cosmetics in the world. The group tucks under its umbrella some of the best known brands and companies in the beauty business: cosmetics houses Lancôme, Vichy and Helena Rubenstein, and fragrance houses Guy Laroche, Cacharel and Ralph Lauren. Given the French penchant for qualitative research, and given the nature of the cosmetics industry, Anne Murray, Head of Research, was asked which type of research she favoured.
‘We’re not particularly pro-quantitative or qualitative. Nevertheless, I do think qualitative in our area is very important. There are many sensitive issues to cover – environmental concerns, animal testing, intimate personal products. And increasingly, we have given to us very technical propositions from the labs, and what is a technical breakthrough to a man in a white coat is not necessarily so to a consumer. So the research department has to be that interface between the technical side and the consumer.’

example
Trend frecasting at Phillips[iii]
Marco Bevolo is in the future business. At Phillips Design, the design department of the electronics multinational, he is responsible for identifying short term trends in popular culture and aesthetics design. Marco believes his job has a lot in common with marketing research. Trend forecasting at Phillips is carried out through a model developed by Marco and his team called CultureScan. The theoretical background of their approach comes from the Birmingham school of popular culture analysis as well as from the kind of cultural analysis performed by scholars who study phenomena such as ‘punks’ in western cities as if they were a tribe in New Guinea. Rather than hinting at tangible design solutions – colour, ‘touch and feel’ or shape – of TVs, MP3 players or other specific products, CultureScan is supposed to provide an insight into the broader, longer term undercurrents in popular culture and aesthetics design all over the world. These broad trends are then customised by Phillips decision teams. CultureScan uses both an internal and external network of experts to collect information on a wide variety of trends. These insights are then filtered by objective tools in order to validate the outcomes and make them actionable. The predictive horizon of CultureScan is 18 to 36 months with the trends refined every two or three years■

In qualitative research, research agencies and companies are continually looking to find better ways to understand consumers’ thought processes and motivations. This has led to a wealth of research approaches, including techniques borrowed from anthropology, ethnography, sociology and psychology. For example, Intel has a specialist team of researchers, including ethnographers, anthropologists and psychologists, whose principal form of research is in the home. ‘People don’t tell you things because they don’t think you’ll be interested. By going into their homes you can see where and how they use their computers,’ says Wendy March, Intel interaction designer of Intel Architecture[iv]
Primary data: qualitative versus quantitative research
As explained in Chapter 4, primary data are originated by the researcher for the specific purpose of addressing the problem at hand. Primary data may be qualitative or quantitative in nature, as shown in Figure 6.1.
Qualitative research
An unstructured, primarily exploratory design based on small samples, intended to provide insight and understanding.
Quantitative research
Research techniques that seek to quantify data and, typically, apply some form of statistical analysis.
Dogmatic positions are often taken in favour of either qualitative or quantitative research by marketing researchers and decision-makers alike. The positions are founded upon which approach is perceived to give the most accurate understanding of consumers. The extreme stances on this issue mirror each other. Many quantitative researchers are apt to dismiss qualitative studies completely as giving no valid findings – indeed as being little better than journalistic accounts. They assert that qualitative researchers ignore representative sampling, with their findings based on a single case or only a few cases. Equally adamant are some qualitative researchers who firmly reject statistical and other quantitative methods as yielding shallow or completely misleading information. They believe that to understand cultural values and consumer behaviour requires interviewing or intensive field observation. Qualitative techniques they see as being the only methods of data collection sensitive enough to capture the nuances of consumer attitudes, motives and behaviour.[v]
[Figure 6.1 near here]
There are great differences between the quantitative and qualitative approaches to studying and understanding consumers. The arguments between qualitative and quantitative marketing researchers about their relative strengths and weaknesses are of real practical value. The nature of marketing decision-making encompasses a vast array of problems and types of decision-maker. This means that seeking a singular and uniform approach to supporting decision-makers by focusing on one approach is futile. Defending qualitative approaches for a particular marketing research problem through the positive benefits it bestows and explaining the negative alternatives of a quantitative approach is healthy – and vice-versa. Business and marketing decision-makers use both approaches and will continue to need both.[vi]
The distinction between qualitative and quantitative research can be in the context of research designs as discussed in Chapter 3. There is a close parallel in the distinctions between ‘exploratory and conclusive research’ and ‘qualitative and quantitative research’. There is a parallel, but the terms are not identical. There are circumstances where qualitative research can be used to present detailed descriptions that cannot be measured in a quantifiable manner, for example in describing characteristics and styles of music that may be used in an advertising campaign or in describing the interplay of how families go through the process of choosing, planning and buying a holiday.
Conversely, there may be circumstances where quantitative measurements are used to conclusively answer specific hypotheses or research questions using descriptive or experimental techniques. Beyond answering specific hypotheses or research questions, there may be sufficient data to allow data mining or an exploration of relationships between individual measurements to take place. The concept of data mining illustrated in Chapter 5 allows decision-makers to be supported through exploratory quantitative research.
The nature of qualitative research
Qualitative research encompasses a variety of methods that can be applied in a flexible manner, to enable participants to reflect upon and express their views or to observe their behaviour. It seeks to encapsulate the behaviour, experiences and feelings of participants in their own terms and context, for example when conducting research on children, an informality and child friendly atmosphere is vital, considering features such as the decoration of the room with appropriately themed posters[vii].
Qualitative research is based on at least two intellectual traditions.[viii] The first and perhaps most important is the set of ideas and associated methods from the broad area of depth psychology.[ix] This movement was concerned with the less conscious aspects of the human psyche. It led to a development of methods to gain access to individuals’ subconscious and/or unconscious levels. So, while an individual may present a superficial explanation of events to themselves or to others, these methods sought to dig deeper and penetrate the superficial.
The second tradition is the set of ideas and associated methods from sociology, social psychology and social anthropology, and the disciplines of ethnography, linguistics and semiology. The emphases here are upon holistic understanding of the world-view of people. The researcher is expected to ‘enter’ the hearts and minds of those they are researching, to develop an empathy with their experiences and feelings.
Both traditions have a concern with developing means of communication between the researcher and those being researched. There can be much interaction between the two broad traditions, which in pragmatic terms allows a wide and rich array of techniques and interpretations of collected data.
Qualitative research is a significant contributor to the marketing research industry, accounting for substantial expenditure (around 16% of all spending on marketing research methods)[x], and is growing. In commercial terms, it is a billion-euro plus global industry. However, it is not just a matter of business value. Qualitative thinking has had a profound effect upon marketing and the marketing research industry as a whole.
Rationale for using qualitative research
It is not always possible, or desirable, to use structured quantitative techniques to obtain information from participants or to observe them. Thus, there are several reasons to use qualitative techniques. These reasons, either individually or in any combination, explain why certain marketing researchers adopt a particular approach (Stage 2 of the marketing research process) to how they conduct research, analyse data and interpret their findings.
1. Preferences and/or experience of the researcher. Some researchers are more oriented and temperamentally suited to do this type of work. Just as some researchers enjoy the challenge of using statistical techniques, there are researchers who enjoy the challenges of qualitative techniques and the interpretation of diverse types of data. Such researchers have been trained in particular disciplines (e.g. anthropology) and philosophies (e.g. hermeneutics) that traditionally make use of qualitative research designs and techniques.
2. Preferences and/or experience of the research user. Some decision-makers are more oriented to receiving support in a qualitative manner. This orientation could come from their training but it could also be due to the type of marketing decisions they have to take. Decision-makers working in a creative environment of advertising copy or the development of brand ‘personalities’, for example, may have a greater preference for data that will feed such ‘artistic’ decisions. In the following example, consider how decision-makers would get to understand and represent the language used by teenagers. Consider also the implications for a brand if marketers do not fully understand the language and values of their target markets.
example
First get the language right, then tell them a story[xi]
Teenagers immediately recognise a communication in their language and are very quick to judge whether advertisers have got it right. They see ads and either like them, reject them, ignore them or, in many cases, discuss them. Teenagers are so fluent in ‘marketing speak’ because marketing and advertising are perceived by them to be the kind of work which can be creative, interesting and acceptable. They discuss with one another the advertising which they perceive to be targeting them.
Pelgram Walters International conducted a study called Global Village. The main contention of the study was that teenagers around the world have a common language, which speaks to them in the filmed advertising medium. Part of the study consisted of focus group discussions of 12- to 18-year-olds. Pepsi’s Next Generation advertisement was criticised by more media-literate teenage markets (Britain, Germany and the US) for stereotyping teens and misunderstanding who they are. The ad was a montage of very hip skateboarding teens, male teens wearing make-up, perhaps implying that Pepsi is for the next generation which looks thus. The main complaint was ‘we don’t look like that’, the teens saying that they were not all the same as one another. By aligning the brand image with these extreme images, the commercial was less appealing to mainstream teen consumers.
3. Sensitive information. Participants may be unwilling to answer or to give truthful answers to certain questions that invade their privacy, embarrass them, or have a negative impact on their ego or status. Questions that relate to sanitary products and contraception are examples of personally sensitive issues. In industrial marketing research, questions that relate to corporate performance and plans are examples of commercially sensitive issues. Techniques that build up an amount of rapport and trust, that allow gentle probing in a manner that suits individual participants, can help researchers get close to participants, and may allow sensitive data to be elicited.
4. Subconscious feelings. Participants may be unable to provide accurate answers to questions that tap their subconscious. The values, emotional drives and motivations residing at the subconscious level are disguised from the outer world by rationalisation and other ego defences. For example, a person may have purchased an expensive sports car to overcome feelings of inferiority. But if asked ‘Why did you purchase this sports car?’ he may say ‘I got a great deal’, ‘My old car was falling apart’, or ‘I need to impress my customers and clients.’ The participant does not have to put words to their deeper emotional drives until researchers approach them! In tapping into those deeper emotional drives, qualitative research can take a path that evolves and is right for the participant.
5. Complex phenomena. The nature of what participants are expected to describe may be difficult to capture with structured questions. For example, participants may know what brands of wine they enjoy, what types of music they prefer or what images they regard as being prestigious. They may not be able to clearly explain why they have these feelings or where these feelings are coming from.
6. The holistic dimension. The object of taking a holistic outlook in qualitative research is to gain a comprehensive and complete picture of the whole context in which the phenomena of interest occur. It is an attempt to describe and understand as much as possible about the whole situation of interest. Each scene exists within a multi-layered and interrelated context and it may require multiple methods to ensure the researcher covers all angles. This orientation helps the researcher discover the interrelationships among the various components of the phenomenon under study. In evaluating different forms of consumer behaviour, the researcher seeks to understand the relationship of different contextual environments upon that behaviour. Setting behaviour into context involves placing observations, experiences and interpretations into a larger perspective.[xii] An example of this may be of measuring satisfaction with a meal in a restaurant. A questionnaire can break down components of the experience in the restaurant and quantify the extent of satisfaction with these. But what effect did the ‘atmosphere’ have upon the experience? What role did the type of music, the colour and style of furniture, aromas coming from the kitchen, other people in the restaurant, the mood when entering the restaurant, feelings of relaxation or tension as the meal went on, contribute to the feeling of atmosphere? Building up an understanding of the interrelationship of the context of consumption allows the qualitative researcher to build up this holistic view. This can be done through qualitative observation and interviewing.
7. Developing new theory. This is perhaps the most contentious reason for conducting qualitative research. Chapter 11 details how causal research design through experiments helps to generate theory. Qualitative researchers may argue that there are severe limitations in conducting experiments upon consumers and that quantitative approaches are limited to elaborating or extending existing theory. The development of ‘new’ theory through a qualitative approach is called ‘grounded theory’, which will be addressed later.
8. Interpretation[xiii]. Qualitative techniques often constitute an important final step in research designs. Large scale surveys and audits often fail to clarify the underlying reasons for a set of findings. Using qualitative techniques can help to elaborate and explain underlying reasons in quantitative findings.
Philosophy[xiv] and qualitative research
Positivist perspectives
In Chapter 2 we discussed the vital role that theory plays in marketing research. Researchers rely on theory to determine which variables should be investigated, how variables should be operationalised and measured, and how the research design and sample should be selected. Theory also serves as a foundation on which the researcher can organise and interpret findings. Good marketing research is founded upon theory and contributes to the development of theory to improve the powers of explanation, prediction and understanding in marketing decision-makers.[xv]
Theory
A conceptual scheme based on foundational statements, or axioms, that are assumed to be true.
Operationalised
The derivation of measurable characteristics to encapsulate marketing phenomena, e.g. the concept of ‘customer loyalty’ can be operationalised through measurements such as the frequency of repeat purchases or the number of years that a business relationship has existed.
The dominant perspective of developing new theory in marketing research has been one of empiricism and more specifically positivism. The central belief of a positivist position is a view that the study of consumers and marketing phenomena should be ‘scientific’ in the manner of the natural sciences. Marketing researchers of this persuasion adopt a framework for investigation akin to the natural scientist. For many, this is considered to be both desirable and possible. A fundamental belief shared by positivists is the view that the social and natural worlds ‘conform to certain fixed and unalterable laws in an endless chain of causation’.[xvi] The main purpose of a scientific approach to marketing research is to establish causal laws that enable the prediction and explanation of marketing phenomena. To establish these laws, a scientific approach must have, as a minimum, reliable information or ‘facts’. The emphasis on facts leads to a focus upon objectivity, rigour and measurement.
Empiricism
A theory of knowledge. A broad category of the philosophy of science that locates the source of all knowledge in experience.
Positivism
A philosophy of language and logic cosistent with an empiricist philosophy of science.
As an overall research approach (using the description of a paradigm or research approach as developed in Chapter 2) qualitative research does not rely upon measurement or the establishment of ‘facts’ and so does not fit with a positivist perspective. However, if qualitative research is just seen as a series of techniques, they can be used to develop an understanding of the nature of a research problem, and to develop and pilot questionnaires. In other words, the positivist perspective of qualitative research is to see it as a set of techniques, applied as preliminary stages to more rigorous techniques that measure, i.e. surveys and questionnaires. This use of qualitative techniques is fine but may be limiting. To conduct in-depth interviews, focus groups or projective techniques, to understand the language and logic of target questionnaire participants makes good sense. However, using qualitative techniques just to develop quantitative techniques can affect how those techniques are used. As an illustration, we will examine how focus groups may be conducted.
Paradigm
A set of assumptions consisting of agreed-upon knowledge, criteria of judgement, problem fields, and ways to consider them.
The term ‘focus group discussion’ is commonly used across all continents, yet it subsumes different ways of applying the technique. There are two main schools of thought, which may be termed ‘cognitive’ and ‘conative’.
1. Cognitive. American researchers generally follow this tradition, which largely follows a format and interviewing style as used in quantitative studies. ‘American-style groups’ is shorthand in Europe for large groups (10 participants on average), a structured procedure and a strong element of external validation. Within the cognitive approach, the analysis or articulation has been worked on before, and so the interviews are largely meant to confirm or expand on known issues.
2. Conative. European researchers generally follow this tradition. This style assumes a different starting point, one that emphasises exploration, with analysis taking place during and after the group. There is less structure to the questions, with group members being encouraged to take their own paths of discussion, make their own connections and let the whole process evolve.
Table 6.1 summarises the differences between the US (cognitive) and European (conative) approaches to conducting focus groups. Note the longer duration of the European approach to allow the exploration to develop. To maintain the interest and motivation of participants for this time period, the interview experience must be stimulating and enjoyable.
Table 6.1 The two schools of thought about ‘focus group discussions’[xvii]
Characteristics
Cognitive
Conative
Purpose
Demonstration
Exploration
Sample size
10–12
6–8
Duration
1.5 hours
1.5 to 6 hours
Interviewing
Logical sequence
Opportunistic
Questions
Closed
Open
Techniques
Straight question, questionnaires, hand shows, counting
Probing, facilitation, projectives, describing
Response required
Give answers
Debate issues
Interviewer
Moderator
Researcher
Observer’s role
To get proof
To understand
Transcripts
Rarely necessary
Usually full
Analysis
On the spot
Time-consuming
Focus of time
Pre-planning
Post-fieldwork
Accusations against other style
‘Formless’
‘Over-controlling’
Suited for
Testing or proving ideas
Meaning or understanding
Output
To be confirmed in quantitative studies
Can be used in its own right to support decision-makers
International marketers have always been aware that qualitative research as it developed in the US and Europe involves quite different practices, stemming from different premises and yielding different results. American-style qualitative research started from the same evaluative premise as quantitative research but on a smaller scale. This made it cheaper, quicker and useful for checking out the less critical decisions. European-style qualitative research started from the opposite premise to quantitative research: it was developmental, exploratory and creative rather than evaluative. It was used as a tool of understanding, to get underneath consumer motivation.[xviii]
The American style uses a detailed discussion guide which follows a logical sequence and is usually strictly adhered to. The interviewing technique involves closed questions and straight answers. This type of research is used primarily to inform about behaviour and to confirm hypotheses already derived from other sources. For this reason, clients who have attended groups often feel they do not need any further analysis; the group interaction supplies the answers. Transcripts are rarely necessary and reports are often summarised or even done away with altogether.
The European style is used primarily to gain new insight; it also works from a discussion guide, but in a less structured way. The interviewing technique is opportunistic and probing. Projective techniques are introduced to help researchers understand underlying motivations and attitudes. Because the purpose is ‘understanding’, which requires a creative synthesis of (sometimes unconscious) consumer needs and brand benefits, analysis is time-consuming and usually involves full transcripts.
In the above descriptions of American and European traditions of applying qualitative techniques, it is clear to see the American perspective being positivist, i.e. aiming to deliver a ‘factual’ impression of consumers. The facts may be established, but they may not be enough – they may not provide the richness or depth of understanding that certain marketing decision-makers demand. So, although a positivist perspective has a role to play in developing explanations, predictions and understanding of consumers and marketing phenomena, it has its limitations and critics. The following quote from the eminent qualitative practitioner Peter Cooper cautions us of what we really mean by the term ‘qualitative’:
‘There is much qualitative research that still hangs on the positivist model or is little more than investigative journalism. Competition also comes from the media with increasing phone-ins and debates described as “research”. We need to be careful about the abuse of what goes under the title “qualitative”.’[xix]
The dominance of positivist philosophy in marketing research has been and is being challenged by other philosophical perspectives, taken and adapted from disciplines such as anthropology and sociology. These perspectives have helped marketing researchers to develop richer explanations and predictions and especially an understanding and a meaning as seen through the eyes of consumers.
Interpretivist perspectives
In general there are considered to be two main research paradigms that are used by marketing researchers.[xx] These are the positivist paradigm and the interpretivist paradigm (though these are by no means the only research paradigms that may be adopted by marketing researchers).[xxi] Table 6.2 presents alternative names that may be used to describe these paradigms.
Table 6.2 Alternative paradigm names[xxii]
Positivist
Interpretivist
Quantitative
Qualitative
Objectivist
Subjectivist
Scientific
Humanistic
Experimentalist
Phenomenological
Traditionalist
Revolutionist
Whilst it may be easier to think of these as quite clear, distinct and mutually exclusive perspectives of developing valid and useful marketing knowledge, the reality is somewhat different. There is a huge array of versions of these paradigms, presented by philosophers, researchers and users of research findings. These versions change depending upon the assumptions of researchers and the context and subjects of their study, i.e. the ultimate nature of the research problem. It has long been argued[xxiii] that both positivist and interpretivist paradigms are valid in conducting marketing research and help to shape the nature of techniques that researchers apply.
In order to develop an understanding of what an interpretivist paradigm means, Table 6.3 presents characteristic features of the two paradigms.
Table 6.3 Paradigm features[xxiv]19
Issue
Positivist
Interpretivist
Reality
Objective and singular
Subjective and multiple
Researcher-participant
Independent of each other
Interacting with each other
Values
Value-free = unbiased
Value-laden = biased
Researcher language
Formal and impersonal
Informal and personal
Theory and research design
Simple determinist
Freedom of will

Cause and effect
Multiple influences

Static research design
Evolving design

Context-free
Context-bound

Laboratory
Field/ethnography

Prediction and control
Understanding and insight

Reliability and validity
Perceptive decision-making

Representative surveys
Theoretical sampling

Experimental design
Case studies

Deductive
Inductive
Comparison of positivist and interpretivist perspectives
The paradigms can be compared through a series of issues. The descriptions of these issues do not imply that any particular paradigm is stronger than the other. In each issue there are relative advantages and disadvantages specific to any research question under investigation. The issues are dealt with in the following paragraphs.
Reality. The positivist supposes that reality is ‘out there’ to be captured. It thus becomes a matter of finding the most effective and objective means possible to draw together information about this reality. The interpretivist stresses the dynamic, participant-constructed and evolving nature of reality, recognising that there may be a wide array of interpretations of realities or social acts.
Researcher–participant. The positivist sees the participant as an ‘object’ to be measured in a consistent manner. The interpretivist may see participants as ‘peers’ or even ‘companions’, seeking the right context and means of observing and questioning to suit individual participants. Such a view of participants requires the development of rapport, an amount of interaction and evolution of method as the researcher learns of the best means to elicit information.
Values. The positivist seeks to set aside their own personal values. Their measurements of participants are being guided by established theoretical propositions. The task for the positivist is to remove any potential bias. The interpretivist recognises that their own values affect how they question, probe and interpret. The task for the interpretivist is to realise the nature of their values and how these affect how they question and interpret.
Researcher language. In seeking a consistent and unbiased means to measure, the positivist uses a language in questioning that is uniformly recognised. This uniformity may emerge from existing theory (to allow comparability of findings) or from their vision of what may be relevant to their target group of participants. Ultimately, the positivist imposes a language and logic upon target participants in a consistent manner. The interpretivist seeks to draw out the language and logic of target participants. The language they use may differ between participants and develop in different ways as they learn more about a topic and the nature of participants.
Theory and research design. In the development of theory, the positivist seeks to establish causality (discussed in detail in Chapter 11) through experimental methods. Seeking causality helps the positivist to explain phenomena and hopefully predict the recurrence of what has been observed in other contexts. There are many extraneous variables that may confound the outcome of experiments, hence the positivist will seek to control these variables and the environment in which an experiment takes place. The ultimate control in an experiment takes place in a laboratory situation. In establishing causality through experiments, questions of causality usually go hand in hand with questions of determinism, i.e. if everything that happens has a cause, then we live in a determinist universe.
Causality
Causality applies when the occurrence of X increases the probability of the occurrence of Y.
Extraneous variables
Variables other than dependent and independent variables which may influence the results of an experiment.
Determinism
A doctrine espousing that everything that happens is determined by a necessary chain of causation.
The positivist will go to great pains to diagnose the nature of a research problem and establish an explicit and set research design to investigate the problem. A fundamental element of the positivist’s research design is the desire to generalise findings to a target population. Most targeted populations are so large that measurements of them can only be managed through representative sample surveys. The positivist uses theory to develop the consistent and unbiased measurements they seek. They have established rules and tests of the reliability and validity of their measurements and continually seek to develop more reliable and valid measurements.
Target population
The collection of elements or objects that possess the information sought by the researcher and about which inferences are made.
Reliability
The extent to which a measurement reproduces consistent results if the process of measurement were to be repeated.
Validity
The extent to which a measurement represents characteristics that exist in the phenomenon under investigation.
In the development of theory, the interpretivist seeks to understand the nature of multiple influences of marketing phenomena through case studies. The search for multiple influences means focusing upon the intrinsic details of individual cases and the differences between different classes of case. This helps the interpretivist to describe phenomena and hopefully gain new and creative insights to ultimately understand the nature of consumer behaviour in its fullest sense. The consumers that interpretivists focus upon, live, consume and relate to products and services in a huge array of contexts, hence the interpretivist will seek to understand the nature and effect of these contexts on their chosen cases. The contexts in which consumers live and consume constitute the field in which the interpretivist immerses themselves to conduct their investigations. In understanding the nature and effect of context upon consumers, the interpretivist does not consider that everything that happens has a cause and that we live in a determinist universe. There is a recognition and respect for the notion of free will.
Case study
A detailed study based upon the observation of the intrinsic details of individuals, groups of individuals and organisations.
The interpretivist will go to great pains to learn from each step of the research process and adapt their research design as their learning develops. The interpretivist seeks to diagnose the nature of a research problem but recognises that a set research design may be restrictive and so usually adopts an evolving research design. A fundamental element of the interpretivist’s research design is the desire to generalise findings to different contexts, such as other types of consumer. However, rather than seeking to study large samples to generalise to target populations, the interpretivist uses theoretical sampling. This means that the data gathering process for interpretivists is driven by concepts derived from evolving theory, based on the notion of seeking out different situations and learning from the comparisons that can be made. The purpose is to go to places, people or events that will maximise opportunities to discover variations among concepts. The interpretivist uses theory initially to help guide which cases they should focus upon, the issues they should observe and the context of their investigation. As their research design evolves they seek to develop new theory and do not wish to be ‘blinkered’ or too focused on existing ideas. The interpretivist seeks multiple explanations of the phenomena they observe and creates what they see as the most valid relationship of concepts and, ultimately, theory. Interpretivists seek to evaluate the strength of the theory they develop. The strongest means of evaluating the strength of interpretivist theory lies in the results of decision-making that is based on the theory. Interpretivists continually seek to evaluate the worth of the theories they develop. A principal output of research generated by an interpretivist perspective should therefore be findings that are accessible and intended for use. If they are found meaningful by decision-makers and employed successfully by them, this may constitute further evidence of the theory’s validity. If employed and found lacking, questions will have to be asked of the theory, about its comprehensibility and comprehensiveness and about its interpretation. If it is not used, the theory may be loaded with validity but have little value.
Evolving research design
A research design where particular research techniques are chosen as the researcher develops an understanding of the issues and participants.
Theoretical sampling
Data gathering driven by concepts derived from evolving theory and based on the concept of ‘making comparisons’.
Summarising the broad perspectives of positivism and interpretivism
The positivist seeks to establish the legitimacy of their approach through deduction. In a deductive approach, the following process unfolds:
Deduction
A form of reasoning in which a conclusion is validly inferred from some premises, and must be true if those premises are true.
§ An area of enquiry is identified, set in the context of well-developed theory, which is seen as vital to guide the researcher, ensuring that they are not naive in their approach and do not ‘reinvent the wheel’.
§ The issues to focus an enquiry upon emerge from the established theoretical framework.
§ Specific variables are identified that the researcher deems should be measured, i.e. hypotheses are set.
§ An ‘instrument’ to measure specific variables is developed.
§ Participants give answers to set and specific questions with a consistent language and logic.
§ The responses to the set questions are analysed in terms of a prior established theoretical framework.
§ The researcher tests theory according to whether their hypotheses are accepted or rejected. From testing theory in a new context, they seek to incrementally develop existing theory.
Such a process means that positivists reach conclusions based upon agreed and measurable ‘facts’. The building and establishment of ‘facts’ forms the premises of deductive arguments. Deductive reasoning starts from general principles from which the deduction is to be made, and proceeds to a conclusion by way of some statement linking the particular case in question.
A deductive approach has a well-established role for existing theory; it informs the development of hypotheses, the choice of variables and the resultant measures.[xxv] Whereas the deductive approach starts with theory expressed in the form of hypotheses, which are then tested, an inductive approach avoids this, arguing that it may prematurely close off possible areas of enquiry.[xxvi]
The interpretivist seeks to establish the legitimacy of their approach through induction. In an inductive approach, the following process unfolds:
Induction
A form of reasoning that usually involves the inference that an instance or repeated combination of events may be universally generalised.
§ An area of enquiry is identified, but with little or no theoretical framework. Theoretical frameworks are seen as restrictive, narrowing the researcher’s perspective, and an inhibitor to creativity.
§ The issues to focus an enquiry upon are either observed or elicited from participants in particular contexts.
§ Participants are aided to explain the nature of issues in a particular context.
§ Broad themes are identified for discussion, with observation, probing and in-depth questioning to elaborate the nature of these themes.
§ The researcher develops their theory by searching for the occurrence and interconnection of phenomena. They seek to develop a model based upon their observed combination of events.
Such a process means that interpretivists reach conclusions without ‘complete evidence’. With the intense scrutiny of individuals in specific contexts that typify an interpretivist approach, tackling large ‘representative’ samples is generally impossible. Thus, the validity of the interpretivist approach is based upon ‘fair samples’. The interpretivist should not seek only to reinforce their own prejudice or bias, seizing upon issues that are agreeable to them and ignoring those that are inconvenient. If they are to argue reasonably they should counteract this tendency by searching for conflicting evidence.[xxvii] Their resultant theory should be subject to constant review and revision.
Ethnographic research
It is clear that an interpretive approach does not set out to test hypotheses but to explore the nature and interrelationships of marketing phenomena. The focus of investigation is a detailed examination of a small number of cases rather than a large sample. The data collected are analysed through an explicit interpretation of the meanings and functions of consumer actions. The product of these analyses takes the form of verbal descriptions and explanations, with quantification and statistical analysis playing a subordinate role. These characteristics are the hallmark of a research approach that has developed and been applied to marketing problems over many years in European marketing research. This research approach is one of ethnographic research.
Ethnography as a general term includes observation and interviewing and is sometimes referred to as participant observation. It is, however, used in the more specific case of a method which requires a researcher to spend a large amount of time observing a particular group of people, by sharing their way of life.[xxviii] Ethnography is the art and science of describing a group or culture. The description may be of a small tribal group in an exotic land or a classroom in middle-class suburbia. The task is much like the one taken on by the investigative reporter, who interviews relevant people, reviews records, weighs the credibility of one person’s opinions against another’s, looks for ties to special interests and organisations and writes the story for a concerned public and for professional colleagues. A key difference between the investigative reporter and the ethnographer, however, is that whereas the journalist seeks out the unusual, the murder, the plane crash, or the bank robbery, the ethnographer writes about the routine daily lives of people. The more predictable patterns of human thought and behaviour are the focus of inquiry[xxix].
The origins of ethnography are in the work of nineteenth-century anthropologists who traveled to observe different pre-industrial cultures. An example in a more contemporary context could be the study of death rituals in Borneo, conducted over two years by the anthropologist Peter Metcalf.[xxx] Today, ‘ethnography’ encompasses a much broader range of work, from studies of groups in one’s own culture, to experimental writing, to political interventions. Moreover, ethnographers today do not always ‘observe’, at least not directly. They may work with cultural artefacts such as written texts, or study recordings of interactions they did not observe at first hand[xxxi], or even as in the following example, the observations of a refrigerator.

example
The Electrolux ‘screen fridge’
Electrolux and Ericsson joined forces to test new products using ethnographic methods. They provided a group of Swedish consumers with ‘screen fridges’ which download recipes from the internet, store shopping lists and have a built-in video camera to record messages. By putting the fridges in peoples homes Electrolux could see how the technology was actually used and find out whether participants would be prepared to pay for it. This type of research is relatively expensive, but it gives in-depth information that could not be generated from a focus group or one-to-one interview. ■

Before developing an understanding of the ethnography in marketing research, it is worth summarising the aims of ethnographic research.
Ethnography
A research approach based upon the observation of the customs, habits and differences between people in everyday situations.
§ Seeing through the eyes o othersf. Viewing events, actions, norms and values from the perspective of the people being studied.
§ Description. Attending to mundane detail to help understand what is going on in a particular context and to provide clues and pointers to other layers of reality.
§ Contextualism. The basic message that ethnographers convey is that whatever the sphere in which the data are being collected, we can understand events only when they are situated in the wider social and historical context.
§ Process. Viewing social life as involving an interlocking series of events.
§ Avoiding early use of theories and concepts. Rejecting premature attempts to impose theories and concepts which may exhibit a poor fit with participants’ perspectives.[xxxii] This will be developed further in this chapter when we examine grounded theory.
§ Flexible research designs. Ethnographers’ adherence to viewing social phenomena through the eyes of their subjects has led to a wariness regarding the imposition of prior and possibly inappropriate frames of reference on the people they study. This leads to a preference for an open and unstructured research design which increases the possibility of coming across unexpected issues.
This final point is illustrated in the following example. This example can be explored in more depth by actually viewing the reviewed film.
example
Kitchen stories[xxxiii]
Kitchen stories was an unlikey contender for one of the best movies of 2004. Kitchen Stories (Salmer Fra Kjokkenet), a Norwegian/Swedish co-production about a 1950s ethnographic study into the kitchen routines of single Norwegian men. Beneath the light humour, the film seeks to make a more serious point about whether the rules governing research stand in the way of reaching a true understanding of people. The film is based upon the real-life story of a team of Swedish researchers who set out to design the perfect kitchen. The researcher, Folke finds himself partnered with the participant from hell; the doddering, grumpy recluse, Isak. Assuming his position on a shaky high chair stuck in the corner of the kitchen, it quickly dawns on Folke that Isak is determined to make his stay as difficult as possible. Both men are under strict instructions not to communicate and in these circumstances, much humour is wrought from Isak’s petty behaviour such as leaving a tap dripping in an attempt to infuriate the well-mannered researcher. Folke’s flouting of the rules mark him out as a less than perfect researcher. The film does not judge Folke but the method he has to work with. The objective nature of the research translates on screen as cold, dispassionate voyeurism. Ethnography in Kitchen Stories is not the science of observing but a clumsy foolish exercise that imposes rules on both men. It is only when Folke climbs down from his ‘pedestal’ that he truly begins to understand Isak. Both the formal, impartility adopted by Folke and Isak’s bizarre behaviour cease, and the two are no longer ‘researcher’ and ‘participant’ – simply friends. ■

The use of ethnographic approaches has rapidly developed in marketing research. Decision-makers are finding great support from the process and findings, as the following example illustrates.
example
Demonstrating the value of air time[xxxiv]
Capital Radio has been looking at the way the radio affects people’s lives. It wanted to show advertisers how listeners relate to the brand and to demonstrate the value of air time. It teamed up with the Henley Centre for a project called Modal Targeting, which highlighted different modes that listeners go through during the day. The intention was to establish when they were most susceptible to certain advertisements. On six occasions they sent researchers to observe listeners for three days at a time. They don’t stay overnight, and they don’t tell the people what they are looking for because they might affect the way they behave. The skill is that they blend into the background and almost become part of the fixtures and fittings. The study showed that there can be a greater difference between how one consumer feels at the start of the day and the end of the day than how two consumers feel at the same time of the day. Advertisers will now be able to be more effective as they know what ‘mode’ listeners are in. They should not be aiming to sell financial products when people are rushing for a train in the morning.■

Ethnography cannot reasonably be classified as just another single method or technique[xxxv]. In essence, it is a research discipline based upon culture as an organising concept and a mix of both observational and interviewing tactics to record behavioural dynamics. Above all, ethnography relies upon entering participants’ natural life worlds – at home, while shopping, at leisure and in the workplace. The researcher essentially becomes a naive visitor in that world by engaging participants during realistic product usage situations in the course of daily life.
Whether called on-site, observational, naturalistic or contextual research, ethnographic methods allow marketers to delve into actual situations in which products are used, services are received and benefits are conferred. Ethnography takes place not in laboratories but in the real world. Consequently, clients and practitioners benefit from a more holistic and better nuanced view of consumer satisfactions, frustrations and limitations than in any other research method.[xxxvi]
A growing trend is for marketers to apply ethnographic methods in natural retail or other commercial environments. There are several objectives that lie behind these studies, one of which is orientated towards a detailed ecological analysis of sales behaviour. In other words, all of the elements that comprise retail store environments – lighting, smells, signage, display of goods, the location, size and orientation of shelving – have an impact upon the consumer experience and their ultimate buying behaviour. The ethnographer’s role is to decode the meaning and impact of these ecological elements. Often, these studies utilise time-lapse photography as a tool for behavioural observation and data collection over extensive periods of time and avoid actual interaction with consumers, as illustrated in the following example.[xxxvii]
example
Top of the Pops[xxxviii]
The point of purchase (POP) is a manufacturer’s last opportunity to have an effect on their customers’ decisions. Awareness of the crucial role of in-store influences is growing, and several POP companies have started offering detailed research on how customers react at the point of sale. To achieve this awareness, Electronic Surveillance of Behaviour (ESOB) gives a detailed understanding of how consumers behave in a shop. Kevin Price, Managing Director of Coutts Design, has formed a partnership with The In-Store Audit to utilise ESOB.
Says Kevin, ‘with ESOB, shoppers are tracked remotely on video around the store and their movements and actions are followed. Because this technique is fairly unobtrusive, we are able to capture natural shopper behaviour as people are not being followed around by a researcher.’ He goes on to explain the complexity of the computer software. ‘The cameras are specially modified and they record a large sample size. They can measure consumer behaviour from entry to exit, following customers around the store and noting the items they touch and the visual cues that they give and get. The cameras may operate for between 10 and 14 days. The information is then analysed and the key clips from the video are used to reinforce the key points that have emerged from the analysis.’
One of the key elements of the above example is the context in which the consumer is behaving. The researcher observes shoppers, taking in and reacting to their retail experience, behaving naturally in the set context. The context of shoppers does not just mean the retail outlet they visit. The processes of choosing and buying products, of using products or giving them as gifts, of reflecting upon and planning subsequent purchases are all affected by contextual factors. Context operates on several levels, including the immediate physical and situational surroundings of consumers, as well as language, character, culture and history. Each of these levels can provide a basis for the meaning and significance attached to the roles and behaviour of consumption.
‘Can we divorce the ways we buy, use and talk about products from the cultural and linguistic context within which economic transitions occur? The answer is an emphatic no‘.[xxxix]
The ethnographer may observe the consumer acting and reacting in the context of consumption. They may see a shopper spending time reading the labels on cat food, showing different brands to their partner, engaged in deep conversation, pondering, getting frustrated and putting tins back on the shelf. They may see the same shopper more purposefully putting an expensive bottle of cognac into their shopping trolley without any discussion and seemingly with no emotional attachment to the product. The ethnographer may want to know what is going on. How may the consumer explain their attitudes and motivations behind this behaviour? This is where the interplay of observation and interviewing helps to build such a rich picture of consumers. In questioning the shopper in the above example, responses of ‘we think that Remy Martin is the best’ or ‘we always argue about which are the prettiest cat food labels’ would not be enough. The stories and contexts of how these assertions came to be would be explored. The ethnographer does not tend to take simple explanations for activities that in many circumstances may be habitual to consumers. Ethnographic practice takes a highly critical attitude towards expressed language. It challenges our accepted words and utterances at face value, searching instead for the meanings and values that lie beneath the surface. In interviewing situations, typically this involves looking for gaps between expressed and non-verbal communication elements. For example, if actual practices and facial and physical gestures are inconsistent with a subject’s expressed attitudes towards the expensive cognac, we are challenged to discover both the reality behind the given answer and the reasons for the ‘deception’.
Ethnographic research is also effective as a tool for learning situationally and culturally grounded language, the appropriate words for everyday things as spoken by various age or ethnic groups. Copywriters and strategic thinkers are always pressed to talk about products and brands in evocative and original ways. Ethnography helps act as both a discovery and an evaluation tool.[xl] To summarise, ethnographic approaches are useful when the marketing research objectives call for[xli]:
1. High Intensity Situations. To study high intensity situations, such as a sales encounter, meal preparation and service, or communication between persons holding different levels of authority.
2. Behavioural Processes. To conduct precise analyses of behavioural processes, e.g. radio listening behaviour, home computer purchasing decisions or home cleaning behaviour.
3. Memory Inadequate. To address situations where the participant’s memory or reflection would not be adequate. Observational methods can stand alone or can complement interviewing as a memory jog.
4. Shame or Reluctance. To work with participants who are likely to be ashamed or reluctant to reveal actual practices to a group of peers. If they were diabetic for example, participants may be reluctant to reveal that they have a refrigerator full of sweet snacks, something that an ethnographic observer would be able to see without confronting the subject.

In these applications, the ethnographer is expected to critically analyse the situations they observe. The critique or analysis can be guided by theory but in essence the researcher develops a curiosity, thinks in an abstract manner and at times steps back to reflect and see how emerging ideas connect. By reacting to the events and participants as they face them, to draw out what they see as important, the ethnographer has the ability to create new explanations and understandings of consumers. This ability to develop a new vision, to a large extent unrestricted by existing theory, is the essence of a grounded theory approach, which is explained and illustrated in the next section.
Grounded theory
The tradition of grounded theory was developed by Glaser and Strauss in the late 1950s and published in their seminal work in 1967.[xlii] At that time, qualitative research was viewed more as impressionistic or anecdotal, little more than ‘soft science’ or journalism.[xliii] It was generally believed that the objective of sociology should be to produce scientific theory, and to test this meant using quantitative methods.[xliv] Qualitative research was seen to have a place, but only to the extent to which it developed questions which could then be verified using quantitative techniques. Glaser and Strauss accepted that the study of people should be scientific, in the way understood by quantitative researchers. This meant that it should seek to produce theoretical propositions that were testable and verifiable, produced by a clear set of replicable procedures. Glaser and Strauss defined theory as:
Grounded theory
Theory derived from data, systematically gathered and analysed.
… theory in sociology is a strategy for handling data in research, providing modes of conceptualisation for describing and explaining. The theory should provide clear enough categories and hypotheses so that crucial ones can be verified in present and future research; they must be clear enough to be readily operationalised in quantitative studies when these are appropriate.[xlv]
The focus upon developing theory was made explicit in response to criticisms of ethnographic studies that present lengthy extracts from interviews or field observations. Strauss sought to reinforce his view of the importance of theory, illustrated by the following quote:
… much that passes for analysis is relatively low level description. Many quite esteemed and excellent monographs use a great deal of data, quotes or field note selections. The procedure is very useful when the behaviour being studied is relatively foreign to the experiences of most readers or when the factual assertions being made would be under considerable contest by sceptical and otherwise relatively well-informed readers. Most of these monographs are descriptively dense, but alas theoretically thin. If you look at their indexes, there are almost no new concepts listed, ones that have emerged in the course of research.[xlvi]
In contrast to the perhaps casual manner in which some ethnographers may be criticised for attempts at developing theory, the grounded theorist follows a set of systematic procedures for collecting and analysing data. This systematic procedure is used to encourage researchers to use their intellectual imagination and creativity to develop new theories, to suggest methods for doing so, to offer criteria to evaluate the worth of discovered theory, and to propose an alternative rhetoric of justification.[xlvii] The most distinctive feature of grounded theory is its commitment to ‘discovery’ through direct contact with the social phenomena under study, coupled with a rejection of a priori theorising. This feature does not mean that researchers should embark on their studies without any general guidance provided by some sort of theoretical understanding. It would be nigh on impossible for a researcher to shut out the ideas in the literature surrounding a particular subject. However, Glaser and Strauss argued that pre-conceived theories should be rejected as they obstruct the development of new theories by coming between researchers and the subjects of their study. In other words, the strict adherence to developing new theory built upon an analytical framework of existing theory, can result in ‘narrow minded’ researchers who do not explore a much wider range of explanations and possibilities. With the rejection of a priori theorising and a commitment to imaginative and creative discovery, comes a conception of knowledge as emergent. This knowledge is created by researchers in the context of investigative practices that afford them intimate contact with the subjects and phenomena under study.[xlviii]
The following example illustrates the use of grounded theory in the development of theory related to the marketing of health visitors. The example is then followed by a description of the process involved in developing theory through a grounded theory process.[xlix]
example
Grounded theory and the marketing of health visitors[l]
Grounded theory was used to conduct a study into the effectiveness of marketing related to health visiting. The collection and analysis of data were conducted simultaneously as the study worked through interviews and observations, all guided through theoretical sampling. There was a line-by-line analysis of interview transcripts with concepts drawn from these to describe events.Categories based on attitudes, behaviour and the characteristics of informants were built through constantly comparing the data in the transcripts and the emerging concepts. The theory that emerged from this process was further guided by a contextual knowledge of the conditions under which any interactions took place, the nature of the interactions between the informants and health visitors, and the consequences of the actions and interactions undertaken by the informants. In addition, memos gathered throughout the process played a crucial part in developing the theory.
The findings and emergent theory centred around the attitudes of the health visitors’ clients. It identified tactics which could enhance ‘selling’ in health visiting, strategies for gaining new clientele and methods for influencing behaviour in order to encourage the prevention of illness. The study addressed the nature of the ‘process’ of health visiting, and the need to promote the service through personal presentations and advertising. Further practical implications emerged from the study by identifying tactics for raising awareness by getting clients to recognise problems early.[li]■
Attempting to gain an objective viewpoint
For the grounded theorist, data collection and analysis occur in alternating sequences. Analysis begins with the first interview and observation, which leads to the next interview or observation, followed by more analysis, more interviews or fieldwork, and so on. It is the analysis that drives the data collection. Therefore there is a constant interplay between the researcher and the research act. Because this interplay requires immersion in the data, by the end of the enquiry the researcher is shaped by the data, just as the data are shaped by the researcher. The problem that arises during this mutual shaping process is how one can become immersed in the data and still maintain a balance between objectivity and sensitivity. Objectivity is necessary to arrive at an impartial and accurate interpretation of events. Sensitivity is required to perceive the subtle nuances and meanings of data and to recognise the connections between concepts. Both objectivity and sensitivity are necessary for making discoveries. Objectivity enables the researcher to have confidence that their findings are a reasonable, impartial representation of a problem under investigation, whereas sensitivity enables creativity and the discovery of new theory from data.[lii]
During the analytic process, grounded researchers attempt to set aside their knowledge and experience to form new interpretations about phenomena. Yet, in their everyday lives, they rely on knowledge and experience to provide the means for helping them to understand the world in which they live and to find solutions to problems encountered. Most researchers have learned that a state of complete objectivity is impossible and that in every piece of research, quantitative or qualitative, there is an element of subjectivity. What is important is to recognise that subjectivity is an issue and that researchers should take appropriate measures to minimise its intrusion into their investigations and analyses.
In qualitative research, objectivity does not mean controlling the variables. Rather it means an openness, a willingness to listen and to ‘give voice’ to participants, be they individuals or organisations. Though this may seem odd, listening is not necessarily a quality that some researchers possess. The following example illustrates the challenges of ‘listening’.
example
Listening is not the same as researching[liii]
According to the Coca Cola Retail Research Group study, the German discount supermarket Aldi is now the strongest retail brand in Europe. Dieter Brandes was an architect of Aldi’s success, he contends that in reaching this position, Aldi never had a ‘grand’ strategy. “We just groped our way forward. It was a dynamic process driven by intuition, incremental adjustments and decisions, whose consequences were not always foreseeable”. This approach is one of being endlessly curious and being confident to take an experimental approach with much thinking and reflection. The essence of this curiosity is listening. Listening is not the same as researching. Research pursues a pre-identified agenda, that of the researcher.Listening gives other people’s agendas top priority. Because it can pick up looming dangers and new opportunities, listening lies at the heart of effective innovation. But it is very hard to do. ■

Thus, good qualitative research means hearing what others have to say, seeing what others do and representing these as accurately as possible. It means developing an understanding of those they are researching, whilst recognising that researchers’ understandings are often based on the values, culture, training and experiences that they bring from all aspects of their life; these can be quite different from those of their participants.[liv] As well as being open to participants, the qualitative researcher reflects upon what makes them, as observers, ‘see’ and ‘listen’ in particular ways. This usually means that, while working on a particular project, the researcher keeps a diary or journal. This diary is used to make notes about the conditions of interviews and observations, of what worked well and what did not, of what questions they would have liked to ask but did not think of at the time. As the researcher reads through their diary in the analysis process, the entries become part of the narrative they explore, they reveal to themselves and to others the way they have developed their ‘seeing’ and ‘listening’. Research diaries will be covered in more detail in examining qualitative data analysis in Chapter 9.
Developing a sensitivity to the meanings in data
Having sensitivity means having insight into, and being able to give meaning to, the events and happenings in data. It means being able to see beneath the obvious to discover the new. This quality of the researcher occurs as he or she works with data, making comparisons, asking questions, and going out and collecting more data. Through these alternating processes of data collection and analysis, meanings that are often elusive at first later become clearer. Immersion in the data leads to those sudden insights.[lv] Insights do not just occur haphazardly; rather, they happen to prepared minds during interplay with the data. Whether we want to admit it or not, we cannot completely divorce ourselves from who we are and what we know. The theories that we carry around in our heads inform our research in multiple ways, even if we use them quite un-self-consciously.[lvi]
Ultimately, a grounded theory approach is expected to generate findings that are meaningful to decision-makers, and appropriate to the tasks they face. As with other interpretivist forms of research, if it is found meaningful by decision-makers and employed successfully by them, there is further evidence of the theory’s validity. Another qualitative approach that is absolutely meaningful to decision-makers in that its primary focus is to deliver actionable results is called action research.
Action research
Background
The social psychologist Kurt Lewin had a main interest in social change and specifically in questions of how to conceptualise and promote social change. Lewin is generally thought to be the person who coined the term action research and gave it meanings that are applicable today.[lvii] In action research, Lewin envisaged a process whereby one could construct a social experiment with the aim of achieving a certain goal.[lviii] For example, in the early days of the Second World War, Lewin conducted a study, commissioned by US authorities, on the use of tripe as part of the regular daily diet of American families. The research question was: ‘To what extent could American housewives be encouraged to use tripe rather than beef for family dinners?’ Beef was scarce and was destined primarily for the troops.
Action research
A team research process, facilitated by a professional researcher(s), linking with decision-makers and other stakeholders who together wish to improve particular situations.
Lewin’s approach to this research was to conduct a study in which he trained a limited number of housewives in the art of cooking tripe for dinner. He then surveyed how this training had an effect on their daily cooking habits in their own families. In this case, action research was synonymous with a ‘natural experiment’, meaning that the researchers in a real-life context invited participants into an experimental activity. This research approach was very much within the bounds of conventional applied social science with its patterns of authoritarian control, but it was aimed at producing a specific, desired social outcome.
The above example can be clearly seen from a marketing perspective. It is easy to see a sample survey measuring attitudes to beef, to tripe, to feeding the family and to feelings of patriotism. From a survey, one can imagine advertisements extolling the virtues of tripe, how tasty and versatile it is. But would the campaign work? Lewin’s approach was not just to understand the housewives’ attitudes but to engage them in the investigation and the solution – to change attitudes and behaviour.
Lewin is credited with coining a couple of important slogans within action research that hold resonance with the many action researchers that practise today. The first is ‘nothing is as practical as a good theory’ and the second is ‘the best way to try to understand something is to change it’. In action research it is believed that the way to ‘prove’ a theory is to show how it provides in-depth and thorough understanding of social structures, understanding gained through planned attempts to invoke change in particular directions. The appropriate changes are in the proof.
Lewin’s work was a fundamental building block to what today is called action research. He set the stage for knowledge production based on solving real-life problems. From the outset, he created a new role for researchers and redefined criteria for judging the quality of the enquiry process. Lewin shifted the researcher’s role from being a distant observer to involvement in concrete problem-solving. The quality criteria he developed for judging a theory to be good, focused on its ability to support practical problem-solving in real-life situations.[lix]
From Lewin’s work has developed a rich and thriving group of researchers who have developed and applied his ideas throughout the world. In management research, the study of organisational change with the understanding and empowerment of different managers and workers has utilised action research to great effect. There has been little application of action research in marketing research, though that is changing. Marketing researchers and marketing decision-makers alike are learning of the nature of action research, the means of implementing it and the benefits it can bestow.
Approach
The term ‘Action Research’ includes a whole range of approaches and practices, each grounded in different traditions, in different philosophical and psychological assumptions, sometimes pursuing different political commitments. Sometimes it is used to describe a positivist approach in a ‘field’ context, or where there is a trade-off between the theoretical interests of researchers and the practical interests of organisation members. Sometimes it is used to describe relatively uncritical consultancy based on information gathering and feedback[lx]. It is beyond the scope of this text to develop these different traditions, so the following describes an approach that is grounded in the Lewin foundations of the approach, and like his work, is applicable to marketing.
Action research is a team research process, facilitated by one or more professional researchers, linking with decision-makers and other stakeholders who together wish to improve particular situations. Together, the researcher and decision-makers or stakeholders define the problems to be examined, generate relevant knowledge about the problems, learn and execute research techniques, take actions, and interpret the results of actions based on what they have learned.[lxi] There are many iterations of problem definition, generating knowledge, taking action and learning from those actions. The whole process of iteration evolves in a direction that is agreed by the team.
Action researchers accept no a priori limits on the kinds of research techniques they use. Surveys, statistical analyses, interviews, focus groups, ethnographies and life histories are all acceptable, if the reason for deploying them has been agreed by the action research collaborators and if they are used in a way that does not oppress the participants.
Action research is composed of a balance of three elements. If any one of the three is absent, then the process is not action research.
§ Research. Research based on any quantitative or qualitative techniques, or combination of them, generates data and, in its analyses and interpretation, shared knowledge.
§ Participation. Action research involves trained researchers who serve as facilitators and ‘teachers’ to team members. As these individuals set their action research agenda, they generate the knowledge necessary to transform the situation and put the results to work. Action research is a participatory process in which everyone involved takes some responsibility.
§ Action. Action research aims to alter the initial situation of the organisation in the direction of a more self-managed and more rewarding state for all parties.
An example of an action research team in marketing terms could include:
§ Marketing researchers: trained in a variety of qualitative and quantitative research techniques, and with experience of diagnosing marketing and research problems.
§ Strategic marketing managers: decision-makers who work at a strategic level in the organisation and have worked with researchers, as well as those who have no experience of negotiating with researchers.
§ Operational marketing managers: decision-makers who have to implement marketing activities. These may be the individuals who meet customers on a day-to-day basis and who really feel the impact and success of marketing ideas.
§ Advertising agency representatives: agents who have worked with strategic decision- makers. They may have been involved in the development of communications campaigns to generate responses from target groups of consumers.
§ Customers: existing customers who may be loyal and have had many years of experience of the company (initiating and funding the action research) and its products and perhaps even its personnel.
§ Target customers: potential customers who may be brand switchers or even loyal customers to competitive companies.
Figure 6.2[lxii] illustrates how action research may be applied. This model of action research is taken from the subject area of the management of change, which is relevant to many of the problems faced by marketing decision-makers. The process aims to create a learning community in a team such as that above. The team develops an understanding of issues to the extent that they make sound judgements and take effective action to implement the changes they wish to make.
[Figure 6.2 near here]
The process in Figure 6.2 can be described as follows.
§ Diagnosis. The present state of affairs would be set out, including the perceived barriers to change and an initial broad statement of desired direction for the organisation. Diagnosis would include documenting the change process and all data gathering activities such as secondary data gathering, surveys, interviews or observations.
§ Analysis. An initial interpretation of data gathered would be made. From this the issues to be tackled would be identified. Summary findings and the development of a framework, with set tasks for team members in subsequent data gathering, would be drawn up.
§ Feedback. Data analyses would be fed back for examination and discussion in the team. ‘Ownership’ of the diagnosis would be developed to formulate a commitment to action.
§ Action. Individual courses of action and the development of broader strategies would be formulated.
§ Evaluation. There would be an ongoing review of methods and outcomes. The effectiveness of any action would be evaluated against agreed criteria and critical success factors.
All of these stages are interrelated, so there is no definitive path that the team would take. In subsequent iterations of activities, the team could move around the stages in any order that suits their needs.
The process is illustrated in the following example where Action Research was used to evaluate a youth Drop-In centre. The detail of the case is limited, but there is sufficient to see Action Research being successfully practiced in an area that could be considered as a marketing research challenge. It also presents the challenge of what other research designs could have worked in these circumstances?
example
Creating ‘The Kit’[lxiii]
A Drop-In Centre for street-involved youth, in a Canadian city, had been running for four years and it was time to evaluate its services. The Centre’s mission and clientele were controversial. Some people felt that a safe place to ‘hang out’ met the initial needs of street-involved youth and allowed staff to reach out informally, build trust, and intervene effectively in crises. Others wanted more structured activities and stricter rules, whilst others thought the Centre attracted ‘high risk’ youth to the area and wanted it shut down completely. The evaluation, identified four objectives:
· To involve youth in designing and implementing an evaluation to measure the impact of Drop-In services
· To improve service delivery to youth.
· To collaborate with community members on long-term solutions to help integrate street youth into the community.
· To make the evaluation instrument available to other youth centres.

A team was built of six youth, two staff members and one outside professional researcher. The group met two afternoons per week at the Drop-In. The evaluation began with the team considering three questions: ‘What do we want to know?’ ‘How will we find out?’ and ‘Who do we need to talk to?’. The team worked together and in small groups to develop the initial questionnaire. Framing questions that would get the information they wanted was a lengthy process, with many drafts and re-drafts. The team moved onto try alternative methods with the youth working together to develop tools that would fit the Centre milieu and engage other street-involved youth. When the tools were ready, each session was advertised in the Drop-In, with pizza as an incentive for participation. An iterative process was established for data analysis, ‘walking through’ responses to open-ended questions to learn the principles and techniques of analysis. The youth worked in pairs to describe and summarise the data. The team produced a formal report, a ‘Kit’ and participated in community and academic presentations about their work. The formal written report was produced by everyone brainstorming the contents, the professional researcher drafting each section, submitting it for feedback to the team and re-drafting. The result was a thorough, well crafted document. The youth took the main role in community and academic presentations. Sharing their expertise publicly helped them gain confidence and pride in their hard work and impressive results. Perhaps the most interesting reporting mechanism was ‘The Kit’, a colourful guide for other youth evaluators. ‘The Kit’ was designed and produced entirely by the youth team members.■

Ethics in marketing research
The researcher and the client must respect participants when conducting qualitative research. This should include protecting the anonymity of participants, not misleading or deceiving them, conducting research in a way not to embarrass or harm the participants, and using the research results in an ethical manner.[lxiv]
The above advice from the Market Research Society contrasts sharply with a controversial approach to retail research that sometimes goes by the name of ‘guerrilla ethnography’ or ‘street research’. This involves observing and talking with consumers in their natural habitats. The researcher in this case commonly does not identify their role as a researcher, nor do they formally state the objectives behind their interaction with consumers. Instead, through the normal course of chatting with fellow customers or sales personnel, an attempt is made to glean information about customer preferences, sales cues and customer language. The benefit here is that the social distance and formal barriers between researcher and subject are broken down and interaction is more ‘natural’ and less subject to contrivance. The main objection expressed by critics is the potential invasion of privacy and somewhat manipulative structure of interaction, and the confusion that may be caused by not being absolutely open to those being observed or questioned.[lxv]
In the mid 1990s, addressing the ESOMAR Annual Conference, Karl Vuursteen, Chairman of Heineken exhorted the virtues of observing consumers. He positively encouraged his managers to visit bars to see consumers enjoying their products in their natural context. This he saw as part of a natural curiosity that decision-makers should have of their customers. Decision-makers naturally meet many of their customers on a day-to-day basis, especially sales personnel, and much is to be learned from these encounters. As long as these encounters respect the anonymity of participants, do not mislead or deceive them, and in no way embarrass or harm them, these encounters are a healthy part of conducting business.
It is up to the researcher to be aware of the harm that participants may suffer and the damage such harm will inflict on all parties involved in an investigation. A reminder of why researchers must be aware of the potential for harm is presented in the following example.
example
The need to pay greater attention to issues of the balance of power, to ensure public cooperation[lxvi]
There are two primary reasons why researchers must be aware of the potential harm to participants and the need to foster a continual willingness to cooperate in research:
1. Qualitative research is doubly dependent on public cooperation. Not only do we rely on participants’ cooperation in taking part, but the non-directive and open-ended nature of qualitative questioning techniques means that we also rely on their being positively engaged by the process, willing and eager to apply their minds, and happy to reveal their thoughts.
2. The nature of qualitative interviews and groups means that moderators will always try to encourage participants to reveal more. This brings with it a responsibility to ensure that safeguards are in place, to deter coercion and avoid the potential danger of abuse of power and control over participants in qualitative research.
Internet and computer applications
The Internet presents huge opportunities for the qualitative researcher. Computer-mediated communication (CMC) can be used to run online versions of semi-structured or in-depth interviews, ‘observation’ of virtual communities, the collection of personal documents, and focus groups. Applying these techniques using the Internet rather than face to face presents a whole array of advantages and challenges.[lxvii] These may be summarised as follows.
Advantages
§ Extending access to participants. Provided that potential participants have access to the technology, researchers can cross time and space barriers. They can reach a much wider geographical span of participants, and also populations that are normally hard to reach, such as people with disabilities.
§ Sensitive subjects. For some participants the sensitivity of the subject being studied may mean that they would not discuss it in a face-to-face manner. The anonymity and distance can help to draw out views from such participants.
§ Interest groups. A variety of online formats, such as chat rooms, mailing lists and conferences, focus on specific topics, drawing together geographically dispersed participants who may share interests, experiences or expertise.
§ Cost and time savings. Issues such as the time and travelling expenses of researchers, the hire of venues and the costs of producing transcripts can make face-to-face interviewing an expensive option for many researchers, especially if they are using qualitative approaches for the first time. The Internet dramatically reduces or eliminates many of these costs and thus makes qualitative approaches more accessible to a wider array of companies and decision-makers.
§ Handling transcripts. As interviews or observations are built up through dialogue on the Internet, many of the potential biases or mistakes that occur through audio recordings can be eliminated.
Challenges
§ Computer literacy for the researcher. Applying qualitative research on the Internet means that some degree of technical expertise is required of the researcher. The extent of expertise depends upon which techniques are being used. For example, moderators of focus groups online will have to learn about the capabilities of the chosen software for running a focus group. They will also have to learn about the specific skills in making the group experience work well online, perhaps by participating in online research conferences or gaining exposure to alternative discussion practices online.
§ Making contact and recruitment. Establishing contact online requires a mutual exchange of email addresses. There is an array of techniques that can encourage potential participants to reveal their address (which will be developed in Chapters 7 and 8) but in essence the researcher has to develop rapport and trust to draw the most out of participants.
§ Interactive skills online. Even if the researcher develops their skills online, it must be remembered that participants use their computing equipment with varying degrees of expertise.
§ Losing access. A key challenge for online studies is to sustain electronic connection with participants for the whole period of the qualitative research process. This is a reminder that, unlike a survey which may be a short, one-off contact with a participant, qualitative techniques may unfold and evolve over time and involve returning to participants as issues develop and theory emerges.
The issues above lay out the broad advantages and challenges of the Internet for the qualitative researcher. As we develop more detailed descriptions and evaluations of qualitative techniques in Chapters 7 to 9, we will examine these points in more detail as there are many rapid innovations occurring to support qualitative researchers.
The impact of the internet has had a profound effect upon how qualitative research is conducted, especially the geographic extent, the nature of data that can be gathered and the sharing and communicating of its analysis. The following example is an illustration of this impact. It is just one of the means that the internet and mobile phone technology can significantly improve qualitative data gathering and analysis.

example
Red Bull re-energising it customer feedback teams[lxviii]
Red Bull was facing an “administrative nightmare”. It’s energy teams, roving groups of employees who meet directly with consumers to discuss the product and collect customer feedback – were having to fax details of their meetings with consumers through to head office. They were getting stacks of paper through that took a lot of time to compile into anything worthwhile. What they wanted was a system that gave them ‘real-time’ results. Blue Trail (www.bluetrail.co.uk) a mobile technology specialist, was chosen to implement a new system. Each of Red Bull’s 30 energy teams (consisting of 4 people each) was given an IPAQ pocket PC which were used to store customer responses. Information was then uploaded via Orange mobile phones to a Red Bull web page, hosted by Blue Trail. The programme drives a workforce that can be used for multiple tasks, including trade audits and customer research. Most recently, the IPAQs were equipped with Nexicam digital camera attachments. The energy teams have always been asked to produce photographic diaries, but the use of the Nexicam enables the images to be used for a fraction of the cost of conventional photography.■

Summary
Qualitative and quantitative research should be viewed as complementary. Unfortunately, many researchers and decision-makers do not see this, taking dogmatic positions in favour of either qualitative or quantitative research. The defence of qualitative approaches for a particular marketing research problem, through the positive benefits it bestows and through explaining the negative alternatives of a quantitative approach, should be seen as healthy, as should the defence of quantitative approaches. Business and marketing decision-makers use both approaches and will continue to need both.
Qualitative and quantitative approaches to marketing research are underpinned by two broad philosophical schools, namely positivism and interpretivism. The central belief of a positivist position is a view that the study of consumers and marketing phenomena should be ‘scientific’ in the manner of the natural sciences. Marketing researchers of this persuasion adopt a framework for investigation akin to that of the natural scientist. The interpretivist researcher does not set out to test hypotheses but to explore the nature and interrelationships of marketing phenomena. The focus of investigation is a detailed examination of a small number of cases rather than a large sample. The data collected are analysed through an explicit interpretation of the meanings and functions of consumer actions. The product of these analyses takes the form of verbal descriptions and explanations, with quantification and statistical analysis playing a subordinate role.
In examining qualitative approaches, ethnography as a general term includes observation and interviewing and is sometimes referred to as participant observation. It is, however, used in the more specific case of a method which requires a researcher to spend a large amount of time observing a particular group of people, by sharing their way of life. The ethnographer is expected to critically analyse the situations they observe. The critique and the analysis can be guided by theory but in essence the researcher develops a curiosity, thinks in an abstract manner and at times steps back to reflect and see how emerging ideas connect. By reacting to the events and participants as they face them, to draw out what they see as important, the ethnographer has the ability to create new explanations and understandings of consumers.
Some ethnographers may be criticised in their attempts at developing theory. In response, the grounded theorist follows a set of systematic procedures for collecting and analysing data. A distinctive feature of a grounded theory approach is that the collection of data and its analysis take place simultaneously, with the aim of developing general concepts, to organise data and integrate these into a more general, formal set of categories.
Ethnographic techniques and a grounded theory approach can be applied in an action research framework. Action research is a team research process, facilitated by one or more professional researchers, linking with decision-makers and other stakeholders who together wish to improve particular situations. Together, the researcher and decision-makers or stakeholders define the problems to be examined, generate relevant knowledge about the problems, learn and execute research techniques, take actions, and interpret the results of actions based on what they have learned. There are many iterations of problem definition, generating knowledge, taking action and learning from those actions. The whole process of iteration evolves in a direction that is agreed by the team.
The Internet presents huge opportunities for the qualitative researcher. Computer-mediated communication (CMC) can be used to run online versions of semi-structured or in-depth interviews, ‘observation’ of virtual communities, the collection of personal documents, and focus groups. Being able to conduct such techniques online offers great opportunities to the qualitative marketing researcher and also many challenges.
When conducting qualitative research, the researcher and the client must respect participants. This should include protecting the anonymity of participants, honouring all statements and promises used to ensure participation, and conducting research in a way not to embarrass or harm the participants.
Questions
1. What criticisms do qualitative marketing researchers make of the approaches adopted by quantitative marketing researchers, and vice-versa?
2. Why is it not always possible or desirable to use quantitative marketing research techniques?
3. Evaluate the differences between a European and an American approach to qualitative research.
4. Describe the characteristics of positivist and interpretivist marketing researchers.
5. In what ways may the positivist and interpretivist view potential research participants?
6. What role does theory play in the approaches adopted by positivist and interpretivist marketing researchers?
7. What does ethnographic research aim to achieve in the study of consumers?
8. Why may marketing decision-makers wish to understand the context of consumption?
9. Describe and illustrate two research techniques that may be utilised in ethnographic research.
10. What stages are involved in the application of a grounded theory approach?
11. Is it possible for marketing researchers to be objective?
12. Why may the Kurt Lewin case of action research be deemed an application of marketing research?
13. Describe the key elements to be balanced in the application of action research.
14. Describe the five interrelated phases of an action research approach.
15. What do you see as the key advantages and challenges of conducting qualitative research using the Internet?
Exercises
1. An advertising agency has selected 3 pieces of music that it could use in a new advertising campaign. It has come to you as a marketing researcher to help in making the case for selecting the right piece of music for their campaign. What would be the case for using qualitative techniques for this task?
2. Would a 12 year old schoolchild, that has not been exposed to any academic theories about ‘service delivery quality’ be more creative and open-minded, and thus better suited to conduct a grounded theory approach, compared to 22 year old Business Studies graduate? Would you view change in any way if the study were about a game or toy that was specifically targeted at 12 year olds?
3. You are a brand manager for Lynx deodorant. You wish to invest in an ethnographic study of young men. Ask another student to play the role of Marketing Director. What case would you make to the Marketing Director about the value of investing in an ethnographic study?
4. In the above case of an ethnographic study of young men for Lynx deodorant, what would you feel to be appropriate contexts or circumstances to conduct this work?
5. In a small group discuss the following issues: “quantitative research is more important than qualitative research because it generates conclusive findings?” and “qualitative research should always be followed by quantitative research to confirm the qualitative findings”

Notes

Figure 6.1 A classification of marketing research data
Figure 6.2 The action research approach

[i] Havermans, J., ‘Research is War’, Research World, (June 2005), 20-21
[ii] McElhatton, N., ‘A research commitment more than skin deep’, Research (May 1994), 10.
[iii] Havermans, J. ‘Trend forecasting applied to fashion and consumption, Research World, (February 2005) 22-24
[iv] Clarke, A., ‘Research takes an inventive approach’, Marketing (13 September 2001), 25
[v] Strauss, A. and Corbin, J.M., Basics of Qualitative Research: Techniques and Procedures for Developing Grounded Theory, 2nd edn (Sage, 1998), 27–34.
[vi] Cooper, P., ‘Consumer understanding, change and qualitative research’, Journal of the Market Research Society 41(1) (January 1999), 3.
[vii] Kenway, J., Keep on moving, Research, (November 2005) p. 36
[viii] Sykes, W., ‘Validity and reliability in qualitative market research: a review of the literature’, Journal of the Market Research Society 32(3), 289.
[ix] De Groot, G., ‘Qualitative research: deeply dangerous or just plain dotty?’ European Research 14(3) (1986), 136–41.
[x] ESOMAR, Industry Study on 2004, Esomar World Research Report, 2005, p. 21
[xi] Flaster, C., ‘First get the language right, then tell them a story’, ResearchPlus (November 1997), 11.
[xii] Gilmore, A. and Carson, D., ‘ “Integrative” qualitative methods in a service context’, Marketing Intelligence and Planning 14(6) (June 1996), 21.
[xiii] Mariampolski, H., Qualitative marketing Research: A comprehensive guide, (California, Sage 2001) p. 24
[xiv] It is recognised that this topic is treated in a superficial manner and is really a basic introduction to the key ideas. Students who are interested in developing these ideas further should see: Hunt, S.D., Modern Marketing Theory (Cincinnati, OH: South Western Publishing, 1991), Potter, G., The Philosophy of Social Science: New Perspectives (Harlow: Prentice Hall, 2000) and Silverman, D., Interpreting Qualitative Data: Methods for Analysing Talk, Text and Interaction (London: Sage, 2001).
[xv] Hunt, S., Marketing Theory: The Philosophy of Marketing Science (Homewood, IL: Irwin, 1983), 2.
[xvi] Hughes, J.A., Sociological Analysis: Methods of Discovery (London: Nelson, 1976), 19.
[xvii] Goodyear, M., ‘Divided by a common language: diversity and deception in the world of global marketing’, Journal of the Market Research Society 38(2) (April 1996), 105.
[xviii] Broadbent, K., ‘When East meets West, quite what does “qualitative” mean?’, ResearchPlus (March 1997), 14.
[xix] Cooper, P., ‘Consumer understanding, change and qualitative research’, Journal of the Market Research Society 41(1) (January 1999), 3.
[xx] Hussey, J. and Hussey, R., Business Research (Basingstoke: Macmillan Business, 1997).
[xxi] See Healy, M. and Perry, C., ‘Comprehensive criteria to judge validity and reliability of qualitative research within the realism paradigm’, Qualitative Market Research: An International Journal 3(3) (2000), 118–26.
[xxii] Hussey, J. and Hussey, R., Business Research (Basingstoke: Macmillan Business, 1997).
[xxiii] Channon C., ‘What do we know about how research works?’, Journal of the Market Research Society 24(4) (1982), 305–15.
[xxiv] Creswell, J.W., Research Design: Qualitative and Quantitative Approaches (Thousand Oaks, CA: Sage, 1994).
[xxv] Ali, H. and Birley, S., ‘Integrating deductive and inductive approaches in a study of new ventures and customer perceived risk’, Qualitatitive Market Research: An International Journal 2(2) (1999) 103–10.
[xxvi] Bryman, A., Quantity and Quality in Social Research (London: Unwin Hyman, 1988)
[xxvii] St Aubyn, G., The Art of Argument (London: Christophers, 1956), 61–75.
[xxviii] Travers, M., Qualitative Research through Case Studies (London: Sage, 2001), 4.
[xxix] Fetterman, D.M. Ethnography: Step by step, (California: Sage 1998), 1
[xxx] Metcalf, P., A Borneo Journey into Death: Berawan Eschatology from its Rituals (Kuala Lumpur: Abdul Majeed, 1991)
[xxxi] Silverman, D., Interpreting Qualitative Data: Methods for Analysing Talk, Text and Interaction, 2nd edn (London: Sage, 2001), 45.
[xxxii] Silverman, D., Interpreting Qualitative Data: Methods for Analysing Talk, Text and Interaction, 2nd edn (London: Sage, 2001), 46
[xxxiii] Tarran, B., Northern exposure, Research, (February 2004) 7
[xxxiv] Steiner, R., “Homing in on consumers”, The Sunday Times, (August 25, 2002) 3.8
[xxxv] Stewart, A., The Ethnographer’s Method, (California: Sage, 1998) 5
[xxxvi] Mariampolski, H., ‘The power of ethnography’, Journal of the Market Research Society 41(1) (January 1999), 79.
[xxxvii] Mariampolski, H., ‘The power of ethnography’, Journal of the Market Research Society 41(1) (January 1999), 84
[xxxviii] Witthaus, M., ‘Top of the Pops’, Marketing Week (21 September 2001), 75–8
[xxxix] Mariampolski, H., ‘The power of ethnography’, Journal of the Market Research Society 41(1) (January 1999), 82
[xl] Mariampolski, H., ‘The power of ethnography’, Journal of the Market Research Society 41(1) (January 1999), 81
[xli] Mariampolski, H., Qualitative marketing Research: A comprehensive guide, (California, Sage 2001) 52
[xlii] Glaser, B. and Strauss, A., The Discovery of Grounded Theory (Chicago, IL: Aldine, 1967)
[xliii] Blumer, H., ‘Sociological analysis and the variable’, in Blumer, H., Symbolic Interactionism: Perspective and Method (Berkeley, CA: University of California Press, 1969), 127–39.
[xliv] Travers, M., Qualitative Research through Case Studies (London: Sage, 2001).
[xlv] Glaser, B. and Strauss, A., The Discovery of Grounded Theory (Chicago, IL: Aldine, 1967), 3.
[xlvi] Strauss, A., Fagerhaugh, S., Suczek, B. and Wiener, C., The Social Organisation of Medical Work (Chicago, IL: University of Chicago Press, 1985).
[xlvii] Locke, K., Grounded theory in management research, (London: Sage, 2001) 33
[xlviii] Locke, K., Grounded theory in management research, (London: Sage, 2001) 34
[xlix] This outline summarises a quite detailed procedure. For a definitive practical guideline to performing grounded theory see: Goulding, C., Grounded Theory: A practical guide for management, business and market researchers, (London: Sage, 2002) and Strauss, A. and Corbin, J.M., Basics of Qualitative Research: Techniques and Procedures for Developing Grounded Theory, 2nd edn (London: Sage, 1998).
[l] De La Cuesta, C., ‘Marketing: A process in health visiting’, Journal of Advanced Nursing, 19 (2), 347-53
[li] Goulding, C., Grounded Theory: A practical guide for management, business and market researchers, (London: Sage, 2002) 86-87
[lii] Strauss, A. and Corbin, J.M., Basics of Qualitative Research: Techniques and Procedures for Developing Grounded Theory, 2nd edn (London: Sage, 1998), 53
[liii] Mitchell, A., The path to success is simple – why do so few stay the course, Marketing Week, (May 19, 2005) 20-21
[liv] Bresler, L., ‘Ethical issues in qualitative research methodology’, Bulletin of the Council for Research in Music Education 126 (1995), 29–41; Cheek, J., ‘Taking a view: qualitative research as representation’, Qualitative Health Research 6 (1996), 492–505.
[lv] Strauss, A. and Corbin, J.M., Basics of Qualitative Research: Techniques and Procedures for Developing Grounded Theory, 2nd edn (London: Sage, 1998), 46–7.
[lvi] Sandelowski, M., ‘Theory unmasked: the uses and guises of theory in qualitative research’, Research in Nursing and Health 16 (1993), 213–18.
[lvii] Greenwood, D.J. and Levin, M., Introduction to Action Research: Social Research for Social Change (Thousand Oaks, CA: Sage, 1998), 17.
[lviii] Lewin, K., ‘Forces behind food habits and methods of change’, Bulletin of the National Research Council 108 (1943), 35–65.
[lix] Greenwood, D.J. and Levin, M., Introduction to Action Research: Social Research for Social Change (Thousand Oaks, CA: Sage, 1998), 19
[lx] Reason, P. and Bradbury, H., Handbook of Action Research, (London: Sage, 2002) xxiv
[lxi] Greenwood, D.J. and Levin, M., Introduction to Action Research: Social Research for Social Change (Thousand Oaks, CA: Sage, 1998), 4.
[lxii] Bate, S. P., ‘Changing the culture of a hospital: from hierarchy to network’, Public Administration 78(3) (2000), 487.
[lxiii] Witmore, E. and McKee, C., Six street youth who could…., in Reason, P. and Bradbury, H., (eds.) Handbook of Action Research, (London: Sage, 2002) 396-402
[lxiv] MRS Code of Conduct (The Market Research Society).
[lxv] Mariampolski, H., ‘The power of ethnography’, Journal of the Market Research Society 41(1) (January 1999), 84.
[lxvi] Murphy, D., ‘Fishing by the Net’, Marketing (22 August 1996), 25.
[lxvii] Mann, C. and Stewart, F., Internet Communication and Qualitative Research: A Handbook for Researching Online (London: Sage, 2000), 17–38.
[lxviii] Case studies, Research in Business, (March 2004) 14

Internal secondary data and the use of databases

2010-06-03T09:02:00.000-07:00

Stage 1 Problem definition
Stage 2 Research approach developed
Stage 3 Research design developed
Stage 4 Fieldwork or data collection
Stage 5 Data preparation and analysis
Stage 6 Report preparation and presentation
Objectives
After reading this chapter, you should be able to:
1. describe the nature and purpose of internal sources of secondary data;
2. describe how different technological developments have increased the array of internally generated secondary data;
3. understand how databases are developing into powerful means to understand consumer behaviour through `electronic observation’;
4. understand how databases support traditional forms of marketing research to build up behavioural and attitudinal `pictures’ of target markets;
5. understand how geodemographic information systems can help in integrating data sources and in the graphical display of findings in a non-statistical manner;
6. describe how the link-up of different databases and survey data can be developed through the use of datawarehouses and be analysed through data mining techniques;
7. understand international data capture issues;
8. understand the ethical problems of having individual consumer data held on databases.

If all you do is stick to research in an organisation like this, then you’re dead really[i].
Overview
Marketing research as a function does not support marketing decision-making in isolation. As discussed in Chapter 1, it is seen by many industry commentators as a component of strategic marketing intelligence. Many information technology and system advances have been made in recent years that have fundamentally changed the way that marketing decisions are supported. For example, significant developments in database technology have meant that scanning systems in retail stores, loyalty card data, store panel data and survey data can be fused together to present very clear and up-to-date `pictures’ of consumers. As well as giving direct support to the marketer, these systems give more focus to marketing research activity and direct support to many stages of research.
This chapter describes how internal secondary data and databases have developed to make major impacts upon how decision-makers are supported. The data collected and analysed through database marketing can be seen as secondary data sources. As with all good secondary data sources, they have a major impact upon the conduct and direction of primary data collection, analyses and interpretation. There are also many ethical issues related to utilising internally generated customer data and the use of databases.
We introduce our discussion with an example of the use of databases. Databases generated within companies or bought in from specialist sources are primarily viewed as a tool to generate direct sales and target promotion activities. However, internally generated customer databases are a secondary data source of value to marketing researchers. Technological developments in the collection, analysis and presentation of data present great opportunities to researchers. There are also potential conflicts with the core premise of anonymity in marketing research. This example illustrates how a customer database can reveal behavioural characteristics of customers, valuable information to decision-makers in its own right. From an understanding of who their customers were and how they were behaving, marketing research could be designed to understand the underlying motives for this behaviour. The database analysis acted as an excellent foundation to a research design in deciding who to research and what issues to focus upon.
example
Database mining meets marketing research[ii]
The Save and Spend loyalty card of DIY retailer Homebase helped to create a database of more than four million active customers and their sales transactions. Homebase could see there were many different types of purchase patterns emerging and asked their marketing research agency for help to understand them. The database told them where people lived and what they were spending, but not things like what sort of house they lived in or their attitudes towards DIY. If Homebase could understand what lay behind patterns of DIY purchasing behaviour for customer types, it could use this information to target its advertising and promotions. Data from store exit surveys were overlaid onto purchasing behaviour from the database. Factors such as the frequency of store visits dominated a base level of segmentation. Statistical modelling then revealed the importance of factors such as distance to store, lifestage and attitudes. This resulted in a clear and actionable segmentation approach. For example, from taking one segment of people looking to decorate their homes in a period style, and overlaying that on the level of disposible income, strong differences were found between men and women. Advertising and promotions could then be tailored to three core types of card-holder: those who dislike DIY and just want value for money; those doing DIY for recreation; and those who want to distinguish their property and hence want the latest gadgets. Homebase calculated that the project enabled them to generate £10 return for every £1 marketing spend■

Internal secondary data
Chapter 4 described the nature, purpose and value of secondary data to marketing researchers. A vital source of secondary data comes from within organisations that commission marketing research, namely internal secondary data. These data are generally seen as being `operational data’, i.e. data that represent the daily activities and transactions of a business. Daily transactions may be held in different departments such as sales, accounts or human resources and stored in different manners. The use of operational data has presented opportunities to researchers for as long as businesses have been recording their daily transactions. Even in the days of transactions being recorded manually, it has been a task for marketing researchers to track down different sources of data and analyse them. Locating and analysing internal sources of secondary data should be the starting point in any marketing research project. The main reasons are that, as these data have already been collected, there are no additional data collection costs, there should be no access problems (though individual managers may make access difficult for personal or political reasons) and the quality of the data should be easier to establish (in comparison with externally generated data).
Most organisations have a wealth of in-house information even if they are not marketing or customer focused, so some data may be readily available. For example, imagine a timber merchant that sells wood to builders and cabinetmakers. It creates invoices for all its sales. Its accounts department handles this process and maintains the data that it generates. Yet, there exists much consumer behaviour data in these invoices. They could be analysed by:
§ What products customers buy
§ Which customers buy the most products
§ Which customers repeat purchases
§ Which customers appear only when there are special offers
§ Where these customers are located
§ How these customers pay – by cash or credit
§ Which customers are the most profitable
§ Seasonal patterns of purchasing behaviour by product types and customer types.
There may also be data that relate to promotions activities such as spending on advertising, trade fairs, sponsorship deals or personal selling. The researcher could look for details of spending in these areas and seek correlations with any of the analyses of customer behaviour. The task facing the marketing researcher is to search for such data, conduct analyses and present these to decision-makers to interpret. Such a process may focus the thoughts of decision-makers, by realising the potential that lies in these data. With this focus, other types of operational data and their value may be realised, managers in other parts of the organisation may release data that they guarded, and connections to sources of intelligence may be generated. Here lies the basis of generating clearly focused primary data collection, and effective marketing research.
Operational data
Data generated about an organisation’s customers, through day-to-day transactions.
More marketing decision-makers have realised the benefits of analysing customer data. This realisation and technological developments in collecting, analysing and presenting customer data have given birth to a concept known as Customer Relationship Management (CRM). The growth of CRM has emerged from the use and subsequent integration of various direct marketing channels (direct mail, telemarketing), and the rise of e-business, e-communication and the increasing use of the internet as a conduit for customer care and sales.[iii] The main challenge of implementing CRM is to integrate customer data from the post, telephone, personal visits and the Internet into a central database so that it holds all transactions and contacts with each customer allied with the ability to update the data constantly and to access it immediately whenever necessary.[iv] An illustration of the use and effect of CRM is presented in the following example.
example
Getting to know clients lifts profits[v]
The principles of CRM are simple. Businesses gather accurate information about customers and prospects. Having identified the customers or segments that account for the highest profits, they devise marketing strategies that differentiate between different groups. Greater resources are focused on higher value customers. Every opportunity is used to amass additional information about each client to personalise sales messages and build a closer relationship.
When Mercury Asset Management began to experiment with CRM, they were able to demonstrate that 59% of their profits came from 1% of their customers. Over six months they moved from having a standard type of literature for all their customers and prospects to no fewer than 7,700 types of literature. Digital printing made this personalisation relatively simple. The first stage of this personalisation process was the compilation of comprehensive information about customers.

Many companies see the benefits of compiling comprehensive information about their customers and invest great amounts in developing and maintaining a customer database. The customer database for many companies is used to drive all marketing strategies. Customer data can be created by companies from past records, promotional devices such as competitions or direct response advertising. The database is used to stimulate marketing activities, and the response from these activities is fed back to improve and update it. Database marketing is a circular activity where every iteration improves the total value of the database.[vi] So, when consumers `hook up’ to an online company, through their PC, their TV or even their mobile phone, they help to develop the customer database. They supply personal details, their choice of products or services and their means of payment. From the knowledge gained from these transactions, new targeted offerings can be formulated, and the nature of the customers’ response can be recorded. As the decision-maker learns more about their customers from transaction data, their awareness of gaps in their knowledge becomes more focused. Where those gaps cannot be filled with transaction data, the marketing researcher plays a vital role in the generation and interpretation of bespoke primary data. In the development of good research design, the customer database can be seen as a resource to the marketing researcher when conducting internal secondary data searches.
Customer database
A database that details characteristics of customers and prospects that can include names and addresses, geographic, demographic and buying behaviour data.
There is a whole array of different means to electronically capture customer transaction behaviour and even potential customers through their search for information to buy services and products. It is beyond the scope of this text to describe the array of CRM technologies, Internet trading and online business. We therefore will just concentrate on a concept introduced in the last chapter, concerning scanned data. From a basis of scanned data we illustrate how other data sources, including primary data from marketing research studies, can be integrated. This serves as a link to examine how decision-makers and researchers make sense of the masses of customer data that may be collected.
Scanning devices
One of the most fundamental technological breakthroughs that has allowed the monitoring of product sales has been the bar code. With scanning devices to read bar codes has come the ability to quickly count and analyse sales. If a new product is launched, scanning data can monitor sales on a daily basis, breaking down the sales by advertising region and the type of outlet. The scanning device is an electronic means of observation. Consumers do not answer any questions, do not identify themselves; they merely enjoy the benefits of supermarket queues moving far more quickly compared with the days of checkout assistants manually entering the prices for individual goods in their baskets.
Scanning device
Technology that reads the UPC code from merchandise by passing it over a laser scanner.
What product bar codes and scanning devices do not do is classify consumers. Classification is fundamental to marketing research, marketing segmentation techniques and ultimately a vast array of marketing decisions. Is the new brand of yoghurt more popular with younger age groups compared with older groups? Have more Calvin Klein shirts been sold to male or female buyers? The following example of the Tesco loyalty card illustrates perhaps one of the most sophisticated means of understanding the characteristics of consumers and of linking their characteristics to their purchasing behaviour. When consumers sign up for a loyalty card, they supply Tesco with personal details that enables them to be classified in a number of manners. Using their loyalty card when making a purchase allows Tesco to make the link between what has been bought with who made the purchase. The example shows what impact this means of electronic observation had upon the performance of Tesco.
Relating customer data to scanning systems
example
Tesco Clubcard[vii]
On the 13th February 1995, the face of retailing in the UK started to change, with the national rollout of Tesco’s loyalty scheme, the Clubcard. It began as an attempt to maintain market share by sustaining and developing the value of its customer base against the then market leader Sainsbury’s and fending off threats lower down the grocery food chain, from price cutters such as Asda. The Clubcard enables it to track £4 out of every £5 spent in their stores and it helped to transform its entire retail strategy, helping it to leapfrog over Sainsbury’s to be the market leader. The Clubcard has 10m active members and generates more than £100m in incremental sales every year. The primary reason for its success is the rich seam of data it provides for the chain. This, says a Tesco spokesman ‘helps us to understand what customers want and develop products and services that are relevant’. The transactional data generated by Clubcard gives Tesco almost unparalleled customer insight. This combined with the intelligence generated from the past use of the data, makes it one of the most successful schemes of its sort. The card is helping Tesco stay well ahead of its rivals and has been instrumental in winning £1 out of every £8 spent nationally and helping them break the £2bn profit barrier.

The essence of the Tesco example is that the scanned data of products sold in their stores is linked to known customers, the system links customer identification to product usage. Any promotional offers, competitive activity, new product offerings, discounts, where to locate a new store to name but a few marketing activities, can be analysed and related to classifications of customer. Many businesses have seen the transformation in the fortunes of Tesco and have tried to replicate their performance through the introduction of their own loyalty card schemes.
The loyalty card is the device that supermarkets, pharmacists, department stores, petrol stations and even whole shopping centres and towns have developed in order to link customer characteristics to actual product purchases.
Loyalty card
At face value, a sales promotion device used by supermarkets, pharmacists, department stores, petrol stations and even whole shopping centres and towns to encourage repeat purchases. For the marketing researcher, the loyalty card is a device that can link customer characteristics to actual product purchases.
The loyalty card may be offered to customers as they make a purchase in a store. They normally complete an application form which may include their name and address, demographic details, household details, media usage, and even some lifestyle characteristics. Once the customer uses their loyalty card, the products they have purchased are scanned and a link can be made through the `swiped’ card to their characteristics that can then be related to the scanned data of their product purchases. In return, the customer earns `points’ for their total spend and may earn additional points for buying particular products. The points gained may be redeemed for cash, additional purchases or even goods and services in other retailers or restaurants.
From the marketing decision-makers’ perspective, many benefits accrue from a loyalty card and product scanning system. The following list summarises the benefits to the marketer.
1. Profiles of customers can be built up. The types of individual that are being attracted to a store can be monitored. The returns and contributions made by particular types of customer can be measured. Profiles of the `ideal’ customer type can be built up, and plans developed to attract that type of customer.
2. Products used and not used. The types of product that are being bought or not bought can be monitored. From the customer profile, other types of product can be added to the range offered. Cross-selling of related products can be undertaken. Linked to the customer profile, actual customer behaviour can be understood more fully.
3. Communications that have worked and not worked. Merchandising displays, money-off coupons, three for the price of two, or a clip-out coupon from a local newspaper, for example, can be linked to individuals and products. The effectiveness of particular types of communication for particular types of consumer can be developed. Reassurance that the customer has made the right decision can be given where the size of purchase warrants it.
4. Distribution methods can be tailored. Certain customer types may prefer the convenience of a small store that they visit more than once a week for small purchases of `staple’ goods. Other customer types may shop once a month for the total household. Retailers can have different shop formats for different customers, may develop home delivery programmes or even develop Internet shopping systems.
The above four factors interact to allow marketing decision-makers to redefine their market(s) and the offerings they make to those markets. The iteration of target market definition and marketing mix tailored to those markets is at the heart of strategic marketing.
From the marketing researchers’ perspective, many benefits also accrue from a loyalty card and product scanning system. The following list summarises the benefits to the marketing researcher:
1. One big laboratory. Experimental methods will be described in Chapter 11 but, in essence, the monitoring of customers, markets and interrelated marketing mix activities allows for many causal inferences to be established. For example, what is the effect, and upon whom, of raising the price of Häagen Dazs ice cream by 10%? What is the effect of inserting a cut-out coupon to give a discount on after-sun lotion, placed in Cosmopolitan magazine?
2. Refining the marketing process. With time series of responses to planned marketing activities, statistical models of consumer response can be built with associated probabilities of a particular outcome. Likewise, models of the consumer over their lifetime can be built. Again, statistical models can be built with associated probabilities of particular types of product being bought at different stages of a consumer’s life.
3. Developing a clear understanding of `gaps’ in the knowledge of consumers. The scanner and loyalty card electronically observes behaviour but does not encapsulate attitudinal data. The nature and levels of satisfaction, what is perceived as good quality service, or what brand image is associated with a particular brand of vodka, are examples of attitudinal data. The use of the database helps to identify target populations to measure and the attitudinal data that need to be collected. In all there can be a much greater clarity in the nature of primary marketing research that tackles attitudinal issues.
4. Linking behavioural and attitudinal data. If attitudinal data are elicited from consumers, the data gathered can be analysed in its own right. It is possible, however, to link the gathered data back to the behavioural data in the database. The term of `fusing’ the data from different sources is used. The key to the fusing lies in identifying individual respondents so that one large dataset is built up. The notion of fusing together databases and survey data from different sources is at the heart of building a strong understanding of consumers. The analytical power that emerges from linking behavioural and attitudinal data has been realised by SPSS as illustrated in the following example.
[Laboratory photo near here]

example
The art of SPSS[viii]
The days of pure data processing are over at SPSS. The company has embarked on a strategy to combine the art of working with massive amounts of transactional data, which provide insight into the behaviour of consumers, with the art of working with attitudinal insights, which are typically derived from marketing research. The combination of database analytics and marketing research results in what SPSS calls Predictive Analytics. The software company decided to take this route when it realised that for decades, the craft of analysing transactional data was wrongly separated from the marketing research profession. In many companies’ headquarters, the departments dealing with data mining and Customer Relationship Management were strictly separated from the other field of providing business intelligence: marketing research. SPSS sees huge opportunities in this convergence both for its clients and for itself. It has started to develop software products that embody this convergence between transactional data analysis and marketing research. They have several products that embody the convergence of database analytics and marketing research, such as Predictive Text Analytics. This application enables companies to automatically process large volumes of answers to open-ended questions, collected through surveys, and combine these with transactional data as well as analyses of a company’s own call centre transcripts .

The benefits listed above show why many marketers and marketing researchers welcome the power of building an iterative customer database, through scanned product purchases and knowledge of customers who make those purchases. There are drawbacks, however, that focus on the nature of the `loyalty card’. Loyalty card schemes may be viewed more as a sales promotion technique in much the same manner as giving trading stamps, a dividend or coupons to be redeemed after a period of saving rather than as a means to capture customer data. Many loyalty schemes fail because they have been founded upon short-term goals and do not capitalise on the consumer insight generated through analysis of the data. Loyalty cards should not be seen as a short-term means to boost sales, as compared with other sales promotion techniques, they incur huge operating costs. Proficient companies that use loyalty cards do so to identify profitable consumers and their needs and tailor the entire shopping experience to them, rather than merely deliver discounts to bargain-hungry fickle consumers.[ix]
There are few questions about the huge costs involved in developing and administering loyalty card schemes. For many retailers, however, the investment allows many marketing and marketing research benefits to be realised. It is only when viewed in the light of offering strategic decision-making power and a complement to an integrated marketing information system that such an investment makes sense. One means of integrating scanned data, with knowledge of customers, and of presenting the relationships and analyses in a spatial manner is through the use of geodemographic information systems.
Geodemographic data
One of the main elements of database power illustrated in the preceding section is the linking of different data sources from both scanner data and customer databases. The ability to create those links and to graphically display analyses has been achieved with the development of geodemographic information systems (GIS). At a base level, a GIS matches geographic information with demographic information, allowing analyses to be presented on thematic maps. This base can be built upon with data from customer databases, databases from other sources, and surveys. The combined data again can be presented on maps and in conventional statistical tables.
Geodemographic information system (GIS)
At a base level, a GIS matches geographic information with demographic information. This match allows subsequent data analyses to be presented on maps.
Thematic maps
Maps that solve marketing problems. They combine geography with demographic information and a company’s sales data or other proprietary information and are generated by a computer.
The geographic dimension is vital as a base to the system. Growing up and living in different geographical locations has an effect upon what we buy and the nature of our lifestyle. Look at the huge diversity of consumers around Europe! It is easy to see differences in consumers and their spending habits, between countries and regions within countries, between cities and towns and areas within a town, and even between different sides of a street. These differences emerge from a variety of factors. The following list summarises the main factors, using extreme examples in places. With closer analysis, more subtle differences can be seen which will be illustrated later in this chapter.
1. Physical geography and climate. Consumers living in hot Mediterranean climates in villages close to the sea may have many different needs and wants compared with consumers in Scandinavian inner cities.
2. Economic history, working opportunities. Consumers who are primarily semi-skilled, working in a declining manufacturing sector, may have many different needs, wants and spending priorities compared with those in a region that attracts recent graduates to work in a burgeoning financial services sector.
3. Political and legal differences. Locations with a history of political and legal domination can affect the types of property and subsequently the types of people who live there. The differences may be national, e.g. with policies that encourage state ownership of property, or tax breaks and discounts so that a rented property may be bought by its tenant. The differences may be regional, e.g. a local council may have structural plans to allow the building of new housing estates for families, on green-field sites on the outskirts of cities.
4. Demographic make-up. Regions made up of consumers living in predominantly retirement areas, such as seaside towns, will have many different requirements from regions that are heavily populated with single young people.
5. Infrastructure links. Infrastructure can include the means of travelling around an area as well as the nature and quality of leisure, sports and shopping facilities. Areas with different levels and quality of infrastructure attract different types of consumer. Families with two cars who can comfortably drive to facilities have different needs and wants when compared with individuals living alone who own a bicycle but not a car.
6. Property types. In different locations particular styles of property may dominate: flats rather than houses, multi-storey rather than low-rise, detached rather than terraced, bungalow rather than house. The type, size, quality and costs of property within an area attract different types of consumer.
Thus, differences can be seen between geographic locations that affect the lifestyle of residents, the array of products and services they buy, their ability to buy different types of products and services, and their hopes, fears and aspirations. The founding premise of a geodemographic information system is that the type of property a consumer lives in says much about their lifestyle and consumption patterns. Property type also encapsulates the other five factors that discriminate between consumers living in different geographic regions. For example, consumers living in small one-bedroom flats over shops in a city centre will tend to have very different lifestyles and consumption patterns from those of consumers living in large detached rural properties. Consumers in different property types have different propensities or probabilities of buying particular goods and services and of undertaking activities that make up their lifestyle. They also have different propensities to use and be exposed to different types of media.
From a marketing decision-making perspective, geography also plays a vital role. Knowing where one’s consumers are located affects the means and costs of distribution. For example, should a retail outlet be built to gain the most returns? Which customers will have to pass our competitors in order to get to us? What features and facilities should the outlet have? The location of consumers also affects the means to communicate with them. Are consumers dispersed over a wide area or tightly clustered together? Do they read the same type of newspaper or magazine? Do they watch the same television programmes or films at the cinema? The following example of Fred Olsen Cruise Lines illustrates how they used geodemographic analyses, it also reveals the links from their customer database to their survey work.
example
Crusing for customers[x]
Founded in 1848, Fred Olsen Cruise Lines (www.fredolsencruises.com) has built a reputation for luxurious voyages in ships that are smaller and more intimate than the norm. With some extremely loyal customers, until recently it required little in the way of marketing to fill the relatively small number of cabins available during the cruising season. This all changed with the launch of Braemar in 2001 and increased competition from P&O, Cunard and Carnival. The company has invested in means to create greater customer insight through its customer database, website and marketing research practices. Geographic analysis has helped to pinpoint locations to target for prospective customers. They have selected catchments around travel agents to target by direct mail. If they run a ‘cruise evening’ then they do targeted mailing to publicise it. Depending upon where customers are located, possible attendees of exhibitions are also selected by location and other segmentation criteria and mailed or emailed with free ticket offers. Their website works directly with their customer database so that segmentation codes and other preference data can be used to target returning customers with cookie-based personalised offers on the home page. In future, customers will be able to update their own preferences, including preferred contact channel, online. Though the company already keys in data from post-cruise questionnaires, another goal of the site is to gather extra personal attributes each time a customer visits and also to perform short surveys. Overall, the site helps to give a much deeper understanding of why customers respond to the Fred Olsen offerings.

A map therefore forms the foundation of a geodemographic information system – a map that can identify all properties in a country, all roads, shopping centres and major facilities in towns and cities. On top of a base map can be laid a range of statistical measures. They typically originate from a number of sources and have the common feature of being able to relate to a specific postcode or zip code. An example of such a system is one produced by Experian (www.experian.com). They have developed systems specifically tailored for a number of countries throughout the world. The sources and details of data available in each of the above countries differ, as does the legislation that determines what can be stored and analysed on databases. Typically for each country, statistics can be gathered and used to develop individual geodemographic information systems based upon census data, postal address files, electoral registers, consumer credit data, directories of company directors, mail order purchase records, car registrations and data on access to retail outlets.
From the data collected, the purpose is to classify consumers on a geodemographic basis. Experian define a geodemographic classification as follows:
Geodemographic classification
This groups consumers together based on the types of neighbourhood in which they live. If a set of neighbourhoods are similar across a wide range of demographic measures, they may also offer similar potential across most products, brands, services and media.
Geodemographic classification groups consumers together based on the types of neighbourhood in which they live. If a set of neighbourhoods are similar across a wide range of demographic measures, they will also offer similar potential across most products, brands, services and media.
With the variables chosen for a particular country, i.e. the types of data that are available to build a geodemographic information system, cluster analyses are performed (Chapter 23 details the nature and purpose of cluster analysis). These analyses help to create consumer classifications, based upon the types of property they live in and the propensity of consumers to have certain lifestyles and behave in particular manners. The analyses ensure that each of the descriptions used is reasonably homogeneous in terms of demographic measurements and consumer behaviour. As well as being able to discriminate and describe distinctive groups of consumers, the analyses have to produce `pictures’ of consumers that are meaningful to marketing decision-makers.
Experian have also produced ‘Mosaic Global’ as a consistent segmentation system that covers over 284 million of the world’s households. It is based on the proposition that the world's cities share common patterns of residential segregation. Each country has their ghettos of Metropolitan Strugglers, suburbs of Career and Family and communities of Sophisticated Singles. In terms of their values and lifestyles each type of neighbourhood displays strong similarities in whichever country it is found. Using local data from 16 countries and statistical methods, Experian has identified 10 distinct types of residential neighbourhood, each with a distinctive set of values, motivations and consumer preferences, which can be found in each of the countries. Mosaic Global uses the data from the national Mosaic classification systems for the following countries: Australia, China (Beijing, Guangzhou, Shanghai), Denmark, Finland, France, Germany, Greece, Hong Kong, Ireland, Netherlands, New Zealand, Norway, Spain, Sweden, UK and the USA.
The resulting analyses have produced a classification of 10 main consumer types. Table 5.1 lists these types and the percentages of each type in the populations of Australia and Sweden. The following example describes characteristics of the Mosaic Global group labeled as Sophisticated Singles.
example
Sophisticated Singles
Sophisticated Singles contains young people, mostly single and well educated, who positively enjoy the variety and stimulation afforded by life in large cities. Typically international in their outlook and with a rich network of personal contacts, they are quick to explore and adopt new social and political attitudes and are important agents of innovation, both in terms of lifestyles and the adoption of consumer products. Most are at the stage of their lives when the development of ‘human’ capital, i.e. skills, contacts, knowledge, continue to take precedence over the maximization of their incomes or indeed the accumulation of financial assets and much of their income is spent on ‘experiences’, such as entertainment, eating out, travel, books and magazines, rather than on equipment. They exhibit a variety of household arrangements and typically marry and have children late in their lives. Such people gravitate towards the smarter downtown areas of major cities where they spend short periods of time living in small, rented apartments.

Table 5.1 MOSAIC Global classification of the Global, Australian and Swedish populations
Classification descriptor
% of Global population[1]
% in Australian population
% in Swedish population
Sophisticated Singles
7.90
10.40
4.34
Bourgeois Prosperity
9.20
22.40
18.61
Career and Family
8.60
3.50
7.13
Comfortable Retirement
2.90
2.00
2.54
Routine Service Workers
9.30
3.80
11.40
Hard Working Blue Collar
10.90
18.90
5.59
Metropolitan Strugglers
18.50
5.00
26.02
Low Income Elders
6.20
4.20
10.25
Post Industrial Survivors
12.20
16.80
2.89
Rural Inheritance
14.60
13.30
11.22

With a geodemographic information system, it is possible to pinpoint where the Sophisticated singles in any of the countries analysed, whether they are clustered in particular regions or cities, or whether they are dispersed. From Table 5.1 it is clear to see differences between countries in the distribution of groups. If one were to dig deeper within each country there would be clear differences in the proportions of these groups between cities and regions. From these classifications and the data that can be added from other databases, models of consumer behaviour can be developed. Customers can be mapped out to see how far they live from a retail outlet or to see whether they pass a competitor’s store to reach a retail outlet. The profile of customers that a company has can be compared with national, regional or city profiles. Data that are captured on customer databases can be mapped out. For example, the ABN AMRO bank can map out which customers have responded to an offer to take out a personal loan at a discounted rate, as well as building up a profile of those who respond. The following example illustrates how Experian’s data and systems are merged with customer databases .
example
Building long-lasting partnerships for Parfümerie Douglas
Parfümerie Douglas is the largest subsidiary of the Douglas Lifestyle Group in Germany. It is the market leader in retail perfumery in Europe, with more than 430 stores in Germany and a further 300 in the Netherlands, Italy, Austria, France, Switzerland, Spain and Portugal. Retail perfumery is a highly competitive market with little customer brand loyalty. The challenge facing Parfümerie Douglas was to maintain its market leadership and achieve its growth ambitions by encouraging greater customer loyalty and increasing average customer spend. Parfümerie Douglas launched its loyalty card in 1995 and by 2001 had issued 1.4 million cards. Since the card's inception, Experian has partnered with Parfümerie Douglas, merging the data generated from the card with their data and systems to provide a full-service account processing function. The Douglas card has met the company's objectives with regard to customer loyalty, sales per customer, payment security and bad debt management. Douglas had their account processing system already in place when German regulations regarding discounts were liberalised in 2001. This gave them an advantage over their competitors by having the ability to test market special offers and discounts through the card's existing database.

In addition to customer behaviour being added to the geodemographic system, survey data can also be added. The key that would link the survey to the customer database may be either a named customer or a postcode. The following example shows illustrates a major survey that utilises the power of Experian’s Mosaic to enable more detailed analyses of target markets. Table 5.2 illustrates and provides links to a number of established and well respected surveys that link with and use geodemographic classifications to generate more insightful analyses of target markets.
example
The MORI Continuous MFS Tracking Survey

The MFS (www.mori.com) continuous tracking survey is widely syndicated to many of the UK's leading banks, building societies, insurance companies and other organisations operating in the personal finance sector. All major markets are measured and the questionnaire is constantly updated to reflect developments in this rapidly changing sector. 2000 different adults are interviewed each fortnightly in-home, face-to-face, by MORI's fieldforce. Quotas are set to reflect the British population by age, sex, class, and region. This large annual sample size of 48,000 enables detailed analyses to be run. The types of questions asked by the MFS Omnibus cover the following topics: Credit cards, Personal Loans, Life & Pensions, Technology Tracker, Current Accounts, General Insurance (Home & Motor, Pet, Travel & Medical), Future Buyer behaviour, Savings & Investments, Loyalty/store Cards. All MFS respondents are post-coded and therefore geodemographic analyses can be run against any group of respondents. This can then be compared to clients' own marketing database for any industry classification.

In this example, an insurance company can map out who bought a new policy from them. They may be able to profile and map out the types of individual who bought different types of financial services beyond insurance. The company may be interested about the levels of satisfaction related to the different financial service companies used by the respondents. The results of the survey can be analysed by the different Mosaic classifications, and levels of customer satisfaction can be mapped out. Additional purchases of insurance related to satisfaction or customer loyalty can be captured. It can be seen from the above that through the use of a geodemographic information system, profiles of target markets, measures of the success of marketing decisions and the means to model consumer behaviour can all be achieved. Graphical representations can be made of customer behaviour, their attitudes and their levels of satisfaction. Using these data, the insurance company additionally has the potential to measure the propensity of potential customers in new locations to buy particular types of insurance policy and other financial services. Using the Mosaic Global classification, the propensity of potential customers could lead to new locations throughout the world.
Table 5.2 Example of marketing research surveys that are linked to geodemographic classifications
Marketing Research survey
Markets covered
Linked Geodemographic classifications
FRS – NOP’s Financial Research Survey www.gfknop.co.uk
Financial
Acorn www.acorn.caci.co.uk
MFS – MORI’s Financial Survey www.mori.com
Financial
Mosaic, www.experian.co.uk CAMEO www.eurodirect.co.uk
TNS – Taylor Nelson Sofres’ Superpanel www.superpanel.tns-global.com
Fast Moving Consumer Goods
Acorn www.acorn.caci.co.uk

Linking different types of data
The previous example illustrated how different types of data can be merged and mapped out to represent customer characteristics. One of the main applications of collecting customer data from different sources and linking it together would be to perform segmentation analyses. Examining the means by which target markets can be segmented, it is clear to see that the five methods as illustrated in Figure 5.1 can be individually utilised or combined to build clearer `pictures’ or profiles of target consumers.
Figure 5.1 gives examples of where data may be obtained from, to help build up profiles of customers and markets. In the example of `psychographics’ or lifestyle measurements, data may be generated from electronic point of sale (EPOS) systems or surveys. In the case of the EPOS collection, the purchasing of particular types of products can indicate characteristics of a lifestyle. In a more direct manner, questions in a survey can help to build a profile of lifestyle behaviour. In its own right, `lifestyle’ can be a valid means of segmenting a market, perhaps positioning products and services to consumers who aspire to a particular lifestyle. However, being able to combine demographic measurements, broader behavioural characteristics and a knowledge of where these consumers live helps to build a `picture’ of consumers that facilitates strong marketing decision-making support.
Figure 5.1 indicates that as one moves from the demographic through to psychological characteristics the measurement process becomes more difficult. Putting aside the differences in techniques to capture `demography’, `behaviour’ or `psychology’, what is being captured becomes more difficult as one moves towards psychological variables. If one considers psychological variables that are vital to marketing which could be captured, examples such as satisfaction, loyalty, trust and quality are not as easy to capture as questions such as gender, age or where one lives. Chapter 12 will explore the concept of measurement in more depth, but at this stage consider what `satisfaction’ actually means, and then the problems of measuring that concept in a valid and consistent manner.
Conversely, as the measurements become more difficult to conduct, they add more to the `picture’ of consumer and market profiles. To say that a market is primarily female, aged between 25 and 40 and lives in a detached property with a mortgage, starts to build a `picture’ of target consumers. To add details of their media behaviour, the array of products and services they buy, characteristics of their lifestyle and their expectations helps to build up a rich and, for decision-makers, very useful `picture’ of target consumers.
Examining the variety of data sources that can be used in the interrelated variables that build market profiles, it is clear to see a role for traditional survey work, scanned data, customer data, externally generated secondary data and the use of loyalty cards. There is a clear interdependence among the different data sources with the increased sophistication of decision support systems that allow the `fusing’ of the data to be conducted.[xi]
[Figure 5.1 near here]
Stages of development in using databases and survey data to build profiles of consumers and model marketing decisions
The last section discussed how different data could be combined to build strong `pictures’ of consumers. Reflecting upon the role of the marketing researcher in supporting the marketing decision-maker as detailed in Chapter 1, it is clear that the combination of survey data and databases plays a major role in fulfilling the following, helping to:
§ describe the nature and scope of customer groups;
§ understand the nature of forces that shape the needs of customer groups and the marketer’s ability to satisfy those groups;
§ test individual and interactive controllable marketing variables;
§ monitor and reflect upon past successes and failures in marketing decisions.
The actual implementation of the decision support systems that allow the combination of data sources to be used in supporting decision-makers can take a great deal of time, expense and organisational learning. It is not the intention here to go through the planning, training and organisational issues in making the systems work, but to broadly summarise the stages that an organisation may go through in combining survey and database data. Figure 5.2 summarises the stages of integration; the following descriptions develop the summarised stages in more detail.
1. Analyse existing consumer database. These data could include the daily operational transactions or enquiries made to a company. As an internal secondary data source it is the cheapest and most readily available data – providing the organisation culture allows access and analysis to marketing researchers.
2. Use supplied geodemographic profiles. There are a growing number of geodemographic systems vendors, some of who have been in operation for over 25 years[xii] In this time they have been able to refine the data they collect and the analyses they produce to build consumer profiles. Companies can buy a base system `off the shelf’ from systems vendors, and add a variety of different databases.
3. Combine existing consumer data with geodemographic profiles. Using the mapping functions of the geodemographic system, existing customer data can be analysed using the profiles supplied with the system. Maps can be used to illustrate the catchment and types of customer and then to evaluate potential in new locations.
4. Use other surveys (either own or from external sources) that build on geodemographic sources and customer database. Surveys conducted by a company where either a customer identification or postcode are recorded can be added. Survey data can be analysed using the geographic profiles and analyses represented using maps.
5. Use combined data sources to create own profiles of customers. Companies gain experience from using the geodemographic profiles and adding their own data. Over a period of time they may see that the generalised definitions of consumers from the geodemographic system do not accurately represent their existing and target customers. With the benefit and use of their own data, they may take the raw data from the geodemographics vendor and produce their own classifications.
6. Datawarehouse analytics. Essentially, this is using many database sources to build one huge database that may be accessed, allowing data to be fused and analysed. The more sophisticated developments of datawarehousing would include the capture and integration of qualitative data that can emerge from intelligence and primary data sources. The analyses that emerge from datawarehouse analytics would be done to suit particular reporting requirements or specific queries from either marketing research or marketing managers. With the growth and significance of this development in decision support, the next section describes the datawarehouse in more detail.
[Figure 5.2 near here]
The datawarehouse
One of the most prolific users and innovators in datawarehouses is the banking industry. The following example illustrates the problems and opportunities for the banking industry of having many different departments with quite distinctive databases.
Datawarehouse
This may be seen as a `super-database’, but more specifically it may be defined as a process of gathering disparate data from database and survey sources, and converting it into a consistent format that can aid business decision-making.
example
You say `warehouse’, I say `database’…[xiii]
Three major changes are sweeping through bank marketing: banks are becoming even more customer driven, they are becoming increasingly information rich, and they are now dependent on constantly evolving computer technologies.
Being more customer driven results in a breaking down of the previously hermetically sealed functional areas of banking. Synergistic marketing and sales is now the name of the game. But as these invisible walls come tumbling down, banks are confronting the unintended result of departments with unique informational needs and personal computers: databases of highly valuable information that have no connection with each other. It is as if the tide has gone out, leaving tidal pools teeming with rich data separate from one another along a beach. In this case, the whole of the data really is greater than the sum of its parts. These pools of customer data are not just valuable in and of themselves. The greatest value is in the across-the-board juxtaposition of all the data pools with one another. That’s where the confusing conceptual model of a datawarehouse comes in. `Datawarehousing’, then, is simply about the creation of a super-database.
The datawarehouse may be seen as a `super-database’, but more specifically it may be defined thus:[xiv]
A datawarehouse is as much a process of gathering disparate data, converting it into a consistent format that can aid business decision-making, as it is a configuration of software and hardware. Datawarehouses empower users by providing them with access to a whole array of information in an organisation, making it available for use in other applications.
From this definition, the datawarehouse can be described as having the following three qualities:
1. It is a collection of integrated databases designed to support managerial decision-making and problem solving.
2. It essentially becomes a giant database that can include survey data held in a database format.
3. It physically separates an organisation’s operational data systems from its decision support systems.
At its most fundamental level, the datawarehouse has three components.
1. Acquisition. This includes all the programs, applications and various interfaces that extract data from existing databases. It continues with preparing the data and exporting it to the datawarehouse.
2. Storage. This is synonymous with any database. It simply involves a storage area to hold a vast amount of data from a variety of sources. The storage area is organised to make it easy to find and use the data. It will be updated from a variety of sources which could be through scanner and loyalty card data on customers, or through the use of intranet data as described in Chapter 4 when examining the compilation of competitor data.
3. Access. This encompasses both set reporting of predetermined events and the means of performing individual analyses, querying `what-if’ scenarios. The process of exploring the databases uses data mining techniques. As marketing researchers and decision-makers learn about markets and their effects upon those markets, the development of predictive models is facilitated. Predictive models are built and tested using historical customer and transactional data, and then used to predict how customers might respond in any number of future situations. As well as the classic marketing application of who may respond to a particular offer, they help to evaluate how likely different individuals or groups are to end their relationship with a company, default on a loan or buy an extra product. The use of predictive modelling started out in credit risk; who is most likely to pay up rather than moving out and disappearing? Analysts found that by weighting different variables such as income appropriately, they could predict with some accuracy the degree of risk that customer presented. Awareness of the power of modelling outside credit and direct marketing rose dramatically with the use of techniques like collaborative filtering by online vendors such as Amazon. If you buy certain books, they you may well be interested in other books bought by customers with a similar history to yours[xv]
Data mining
The process of discovering meaningful correlations, patterns and trends by sifting through large amounts of data stored in repositories, using pattern recognition as well as statistical and mathematical techniques.
Data mining
Data mining is a process of discovering meaningful correlations, patterns and trends by sifting through large amounts of data stored in repositories. The process uses pattern recognition as well as statistical and mathematical techniques.[xvi] Data mining should not be confused with datawarehousing. The datawarehouse could be termed a `repository’ or a place where large amounts of sometimes disparate sources of data are stored; data mining is a process that depends upon access to the data held in that repository.
Examples of what data mining aims to do are as follows.
§ Classify customers into specific categories that are meaningful to decision-makers
§ Identify potential target markets that possess the characteristics that decision-makers seek
§ Forecast sales or the use of services
§ Discover which types of products or services are purchased together
§ Discover patterns and trends over time, such as `after graduation, students take a holiday’, and be able to show the probabilities associated with different holiday types.
Data mining is a way of exploiting the data held by organisations to help discover and develop specific information or knowledge. As well as using proprietary software to perform pattern recognition, statistical and mathematical techniques, it can also be seen as a mental process undertaken by decision-makers. The decision-makers who interact with large datasets using data mining are generally not specialists in statistics, data analysis, datawarehousing and other data tools; they are information users, seeking support for their decision-making processes. With the aid of data mining software, the decision-maker is encouraged to think in new ways and ask new questions. The decision-maker discovers more from the data and explores in new areas, integrating other sources of data, going through iterative processes to dig deeper. The exploration process involves the discovery of non-trivial relationships of dependence or associations, non-trivial clusters, factors or trends and an understanding of the managerial significance of these discoveries. A data mining process must be `user- oriented’ and that user is typically the decision-maker. The following example illustrate the processes described above by showing how data mining has been used to explore the customer database of one of Europe’s leading banks.

example
Credit Suisse
One of the world's leading financial services companies, Credit Suisse Group provides banking and insurance solutions for private clients, companies and institutions. Based in Zurich, Switzerland, Credit Suisse employs 80,000 people worldwide. Competition in the financial services industry is intense, and obtaining new customers is expensive. In order to maximize profitability, Credit Suisse focused on three areas. First, identifying profitable current customers. Second, managing customer relationships to ensure longevity. Third, retaining profitable customer accounts. In 1997, Credit Suisse started the "Loyalty Based Management" program, with the primary goal of retaining profitable customers. They invested in a six-member "data mining team" that used the Clementine tool (www.spss.com/clementine) to analyse a data warehouse of its 2.5 million customers, each with more than 400 attributes. The analysis was used to identify potential leads among Credit Suisse's customers and intelligently market to them based on their individual preferences and histories. With the help of SPSS, Credit Suisse's data mining activities, analysis and modeling, have been fully integrated into their business processes and have proven their value in many different applications. As a result of the success of the Loyalty Based Management project, Credit Suisse consultants began to see data mining's benefits and started to use it to sell specific customers targeted services. Credit Suisse can now identify customers, typically the top one percent, who are extremely likely to buy a service, thus increasing the opportunities for cross-selling and retaining customers. Detailed segmentation of its vast customer base allows Credit Suisse to develop targeted solutions for its customers. This segmentation is executed inductively using the cluster algorithm and the dimensions are tailored directly to the customer requirements. Each cluster serves as a starting point for individual marketing campaigns. This hierarchical system is advantageous because the customer database is continually researched and monitored. As a result, changes in the cluster structure are quickly identified and appropriate responses are triggered. ■
Databases and marketing research
There has been a phenomenal growth in the use of databases to support marketing decision-making. In larger organisations with many divisions or where mergers and acquisitions have taken place, the datawarehouse has facilitated the `fusing’ of data from many sources. Such developments are seen as a threat by many in the marketing research industry. However, many marketing research companies and marketing research departments within companies are embracing database techniques, utilising the synergistic benefits of matching database analyses with traditional survey data through data mining. To illustrate this point, consider the following quote from Greg Ward, Development Director for Taylor Nelson, (www.tns-global.com) the largest marketing research company in the UK.
The marketing research industry needs to acknowledge that databases are serious products and that both types of information have benefits. If you take the best of both – what we call information based marketing – you get something that is significantly more powerful. The `them and us’ situation does nobody any favours and the idea that the two disciplines bear no resemblance to each other is wrong.
As marketing decision-makers become more willing and able to interrogate databases and to creatively generate their own decision support, this does not mean the end of `traditional’ marketing research. As illustrated earlier when examining types of data that are used to build consumer and market profiles, psychological data play a vital role that is fulfilled by qualitative and quantitative marketing research. The marketing researcher needs to develop a greater awareness of both how data captured through traditional methods can be integrated with data held on databases and how the combined data creatively support decision-makers.
Databases, the development of datawarehouses and the use of data mining techniques are allowing a wider and shared use of data. The graphical formats of presenting data, especially using maps, break down many barriers in decision-makers who resist formal statistical analyses. They encourage managers to tailor output to meet their individual needs. The creativity that is the hallmark of marketing decision-making is supported by the creative collection and connections between data. Where there are gaps in decision-makers’ knowledge, they can be more focused and precise in determining what marketing research support they need. Many marketing researchers are rising to meet this challenge. [xvii]
The next example illustrates how data mining can provide in-depth analyses of survey data at great speed or it can help combine research data with other types of data to provide a much richer source of insight. It illustrates how companies can integrate databases and traditional marketing research, seeing the disciplines as complementary, not competitive.

example
TrustMark CFI
TrustMark CFI, part of the worldwide CFI Group, is a market research and management consulting business in Zollikon, Switzerland. In a project for a banking customer, TrustMark CFI consultants used Clementine (www.spss.com/clementine) to analyse customer satisfaction data with operational data, and then build models that would forecast future consumer activity. By combining the two data sources, the models produced by Clementine were more accurate—and comparing the predictions from the Clementine model with reality proved this. When using operational data only, Clementine correctly identified 65.4 percent of the customers who had increased their asset volume significantly, and 56.9 percent of those who had greatly reduced it. However, using the two data sources together, the model was correct in 67.9 percent and 60.7 percent of cases respectively. According to Leonie van de Vijfeijken, a research analyst at TrustMark CFI, the results would have been even better if there hadn't been a large time lag between gathering the operational and survey data. Although the difference in the figures doesn't sound remarkable, it is significant to the bank's business and helps reduce its marketing costs," Using these models on current data allows the bank to predict whether an account is likely to be profitable, at risk, or inactive. This information enables it to target marketing resources more effectively and, in the final analysis, improve financial performance. Interestingly, had the bank not used predictive models and simply run standard analyses on the survey research data, a very different picture would have emerged. When past research data was compared with reality, it was shown that only 21 percent of the respondents who said they were going to increase their asset volume actually did so. ■
There is a blurring of the lines between traditional marketing research and other corporate intelligence activities, including data mining. Whilst some research departments are reinforcing the line that separates the Marketing Research Department from other functions, many are embracing the change. The reason that this integration matters is that it should be the case that the Marketing Research Department is in the best position to drive the understanding of the ‘why’ of behaviour, rather than just describing the behaviour.[xviii] Given the different but complementary roles that marketing research and database analytics perform, put together there can be a great synergistic effect. If a company has a customer transactional database, it’s very unlikely that it can tell them why their customers use the products they do and what media they consume. What marketing research can do is overlay those attitudes, identify segments and make the whole database actionable. [xix]

International marketing research
Linking databases and survey data is transforming international marketing research. Within individual companies, customers may be analysed from operational data within a country, showing different patterns of behaviour between different regions or cities and relating that behaviour to their marketing activities. When a company operates across borders, country differences become just another geographical variable.
In deciding to operate or develop in a particular country, companies may buy a geodemographic information system (should one be developed for that country). A GIS may be used as a foundation to add to their operational data. From this point they may go through the stages of database development as laid out in Figure 5.2. However, international marketers face a dilemma in choosing between ‘consistency’ and ‘richness of detail’
It is possible to create geodemographic systems tailored to individual countries across the globe based upon the data sources available within that country, its data protection laws and how current and accurate those data sources are. Such systems work extremely well if ones marketing activities are based within that country. The problem arises when as an international marketer, one wishes to have consumer classifications that are consistent across borders. Mosaic Global is one of many solutions to this problem. However, in resolving the problem of consistency across countries, an amount of the richness of individual country detail may be lost. To illustrate this dilemma, examine the following summaries of MOSAIC Global, UK and London. See how as the system becomes more focused geographically, there is a greater richness of detail but a loss of consistency in the measurements and classification across borders
example
MOSAIC classifications in London, the UK and the Globe

Mosaic Global is a consistent segmentation system that covers over 400 million of the world’s households. Using local data from 25 countries, Experian has identified ten types of residential neighbourhood that can be found in each of the countries, each with a distinctive set of values, motivations and consumer preferences. Mosaic Global is based on a simple proposition that the world's cities share common patterns of residential segregation. In terms of their values and lifestyles, each type of neighbourhood displays strong similarities in whichever country it is found.
Mosaic UK is the latest version of the market-leading consumer segmentation product. It classifies all 24 million UK households into 11 groups, 61 types and 243 segments, and is updated each year. A development team of over 30 staff took over two years to build Mosaic UK. The result is a classification that paints a rich picture of UK consumers in terms of their socio-demographics, lifestyles, culture and behaviour, providing the most accurate and comprehensive view of UK society at the start of the 21st century.
Mosaic London is a geodemographic classification of the London. The classification covers 5.9 million households (25 % of Great Britain) in approximately 380,000 postcodes. Mosaic London describes Londoners in depth: each postcode has been allocated to one of 41 types and 12 groups that are specific to London. Groups of people are identified that are important to the London community but may not occur elsewhere in the country■
Performing analyses within countries can be most fruitful, provided a base GIS has been established. Problems start where there is no base GIS. In many countries there are great problems in tracking down and combining data sources that can be relied upon. Further, even if reliable data can be located, legislation may make the use of certain data types illegal. In many developing countries, the data needed to build a GIS are sparse. With the data that are available, much experimentation is needed to enable valid classifications that reflect consumer types which are useful to marketers and marketing researchers.
Ethics in marketing research
Marketing researchers are confronted by problems posed by the wording of ESOMAR and individual country marketing research associations’ codes of conduct.[xx] The codes specify that the compilation of lists, registers or databanks of names and addresses for any non-research purpose shall in no way be associated directly or indirectly with marketing research. The view is that whilst marketing research and marketing database analytics are complementary, because the two activities are subject to different legislative requirements and in the interests of transparency for respondents, the ICC/ESOMAR International Code states ‘When acting in their capacity as researchers, the latter must not undertake any non-research activities, for example database marketing involving data about individuals which will be used for direct marketing and promotional activities. Any such non-research activities must always, in the way they are organized and carried out, be clearly differentiated from marketing research activities’. For more details see www.esomar.org/codesandguidelines.
However, the examples detailed in this chapter show that supporting marketing decision-makers through databases and marketing research can be seen as part of a total information industry. Evidence of the many leading marketing research agencies involved in data collection and analyses through databases illustrates that databases need not be unethical. With due care it is possible to combine marketing research ethics and databases generated through database marketing. There are a growing number of companies that have used marketing research for many years that now combine the traditional role of marketing research manager with a wider role including database management. An essential part of this combined role lies in the management of customer databases, adding survey details to respondents’ individual details, at either an individual or an aggregated level.
Given the phenomenal growth of databases in marketing and the support they offer to marketing decision-makers, they are here to stay. With well-planned `traditional’ marketing research integrated into database analyses, the strategic power of consumer and market analyses is phenomenal. If marketers abuse their knowledge of consumers, they stand to do great harm to their brands and corporate image. For example, in bank databases there are many opportunities for the cross-selling of products. Rather than welcoming the approach from another division of a bank, trying to sell insurance to an investment client, there can be a reaction against the approach, affecting the original business. Consumers are now more aware of how valuable knowledge of their behaviour is and how it is used by marketers. They are willing to trade this knowledge for the kind of rewards that are gained from the use of their loyalty cards. Marketers are aware of the dangers of abusing the knowledge that their customers impart to them. However, there are issues of civil liberties that cannot be ignored. These are touched upon in the following example.
example
Loyalty for sale[xxi]
Provided that shoppers like the benefits and do not object to a system which records every bar of chocolate and bottle of gin purchased, no great harm will be done. However, there is a danger that, despite the safeguards of Data Protection Acts, this mass of information on consumers’ habits could leak across the networks into unscrupulous hands. Issues of civil liberty would be raised if, for example, insurance companies could use the data to identify people whose purchases indicated an unhealthy lifestyle; or if the police could draw up a list of suspects by monitoring the purchase of specific items or unusual consumption patterns.
One of the benefits of the use of geodemographic systems is that in many cases the individual does not have to be identified; the postcode is a sufficient key to make a link between databases. This maintains the marketing research industry’s maxim of respondent confidentiality
Internet and computer applications
At the start of the 1990s, this chapter would not have existed in a marketing research text. The idea of conducting internal secondary data searches and analyses would have merited a paragraph or two as part of the process of developing primary data collection. Since then, the massive technological changes that have made the global use of the internet commonplace, the increased storage space, speed and analysis capabilities of computers, and the increased sophistication of software collectively have made fundamental changes to the environment in which marketing researchers operate. The collection and analysis of customer data held within businesses have developed enormously and marketing researchers cannot ignore these developments.
In Chapter 1 we discussed how, for many years, marketing researchers have recognised the competition they face from an array of management consultants, but more recently competition has emerged from raw data providers such as call centres, direct marketing, database marketing and telebusinesses.[xxii] Much of the new competition has emerged from organisations that have utilised and developed Internet and computer applications and have been able to offer support to decision-makers, in faster, cheaper and more user-friendly formats, though not necessarily more rigorously or ethically. Leading marketing researchers have developed means of integrating many of the new formats for supporting decision-makers with traditional marketing research methods and have maintained their customary rigour and ethical behaviour. In order to get a feel of how decision-makers may be supported by some of the means discussed in this chapter, we recommend that you explore the following Websites. These Websites contain case studies of the applications of decision support. When working through these cases, consider what `gaps’ may still exist in the knowledge of decision-makers that may be filled only by the use of traditional marketing research techniques. The following example illustrates such a case. It shows how SPSS are using the integration of different data sources to build consumer behaviour models and predict behaviour. The SPSS concept of Predictive Analytics was presented earlier in the chapter; this example illustrates its application in combining data mining and marketing research (www.spss.com).
example
Top Dutch Insurer Interpolis Selects SPSS Software for Lead Generation
Interpolis, (www.interpolis.com) is the largest single brand consumer property and casualty insurance company in the Netherlands. The company is also active in Luxemburg and Ireland. It has about 6,000 employees, and realized a turnover of 5 billion Euro in 2004. Interpolis have purchased SPSS’ predictive analytics software to further grow its business through lead generation. Harnessing SPSS’ PredictiveMarketing, Interpolis want to predict which customers are most likely to buy its insurance products. With SPSS, Interpolis aims to increase sales to its existing customers and broaden its customer base, which can amount to millions of Euros in additional revenue. Interpolis generates 85 percent of its revenues from Rabobank’s branch network. The insurer selected SPSS’ analytics software to better understand and predict customer needs and feed this information to branch staff, providing a clear indication of opportunities to sell its insurance products. Interpolis’ marketing department will also utilize SPSS’ software to identify which customers are most likely to defect once an insurance contract has ended, the reasons for defection, and the best actions to minimize it. “Interpolis is a great example of a predictive enterprise – leveraging and acting upon customer knowledge across multiple channels: at its marketing campaigns, in branch offices, and in the call center,” added Marcel Holsheimer, Vice President of Marketing for SPSS’ Platform and Applications. “Deploying predictive analytics gives it the ability to improve across several business processes and considerably enhance bottom line performance”.
n

Two major suppliers of software that help with customer relationship management mentioned at the start of this chapter are SAP www.netsuite.com and AIT www.ait.co.uk. They also show how operational data generated in organisations may be integrated. The main geodemographic information systems to examine individual countries and across boundaries were listed earlier in the chapter, but as a reminder, they were: Acorn (www.acorn.caci.co.uk), Mosaic, (www.experian.com) and CAMEO www.eurodirect.co.uk. Numerous datamining software programs are now available. Kdnuggets (www.Kdnuggets.com) is an online resource for dataminers where it has identified the most popular datamining software packages. The most popular vendors and their datamining software programs include the following:
SPSS (www.spss.com/clementine), SAS (www.sas.com/technologies/analytics),
IBM (www-306.ibm.com/software/data/db2bi), Angoss (www.angoss.com),
Megaputer (www.megaputer.com)
Each of these programs can run on client-server platforms and include a wide range of analysis tools and techniques, including neural networks and decision trees[xxiii]. A French datawarehouse supplier with a case based on the Renault Company can be found on www.decisionnel.fr. Finally, visit www.crisp-dm.org to see an organisation devoted to the development of data mining methodology.
Summary
The overall tone of this chapter has been to demonstrate that internally generated secondary data offer great opportunities not only for decision-makers but also for researchers. As with all good secondary data sources, they have a major impact upon the conduct and direction of primary data collection, analyses and interpretation.
Databases, including customer operational data, geodemographic data and survey data, are radically changing how marketing decision-making is being supported. There is much debate as to whether the use of databases is compatible with traditional techniques of marketing research. With the junk mail connotations of databases and compromises of respondent anonymity, many marketing researchers may seek to keep the marketing database at arm’s length. However, handled with the professional acumen that marketing researchers have displayed for many years, the database presents great opportunities for the marketing researcher. In Europe, many of the leading marketing research agencies and research functions within companies have embraced the marketing database as an essential component of their desire to create ‘customer insight’.
For the marketer, databases help to build profiles of consumers, linked to the products, communications and distribution methods those consumers favour. For the marketing researcher, databases can present the opportunity to experiment in `one big laboratory’, build models of consumer behaviour, develop an understanding of the gaps in knowledge of consumers and make links between behavioural and attitudinal data.
Much of the data that offers these benefits has been gained using data that capture customer buying behaviour. The use of the `loyalty card’ is one example. Different types of data, including scanner data, loyalty card data and survey data, may be combined using geodemographic information systems (GIS). Using base geographic and demographic data, existing customers can be analysed and mapped out.
Disparate database sources are pulled together through the use of datawarehouses. The datawarehouse integrates databases and survey data, allowing creative connections between data to be explored. Integrated databases and survey data can be explored using data mining techniques. These processes involve the use of proprietary software and inquisitive, creative decision-makers who search for trends and patterns in customer behaviour and attitudes. The development of datawarehouse and data mining expertise especially helps to cope with the problems of disparate databases and survey data from different countries.
The ethics of using databases provokes much debate in the marketing research industry. As many research practitioners grow more accustomed to using databases, marketing research guidelines and codes of practice are being developed to reflect the good practices that exist in many companies.
Questions
1. How may `operational data’ held in organisations help to build up an understanding of customer behaviour?
2. What is a customer database? Why may a marketing researcher wish to analyse the data held in a customer database?
3. What kinds of data can be gathered through electronic scanner devices?
4. What other sources beyond electronic scanner devices electronically observe customer behavior?
5. Describe the benefits to the marketing decision-maker of being able to capture data that identifies characteristics of consumers and their shopping behaviour in a store.
6. Describe the benefits to the marketing researcher of being able to capture data that identifies characteristics of consumers and their shopping behaviour in a store.
7. Why may the characteristics of consumers differ, based upon where they live?
8. What is a geodemographic classification of consumers?
9. How can the graphical representation of consumer characteristics using maps help marketing decision-making?
10. What benefits may be gained from fusing together customer characteristics held as internal secondary data, with a proprietary geodemographic information system held as external secondary data?
11. How does the compilation of different types of data help to build a strong `picture’ of consumer characteristics?
12. Describe the stages of development in using databases and survey data to build profiles of consumers and model marketing decisions.
13. What is a datawarehouse?
14. What is the difference between a datawarehouse and data mining?
15. Why may there be a difference between a individual country geodemographic classification and a European or Global classification of that same country?
Exercises
1. Visit the websites of Acorn (www.acorn.caci.co.uk), Mosaic, (www.experian.com) and CAMEO www.eurodirect.co.uk. Imagine that you have been commissioned to select a geodemographic system to help a newspaper publisher in a major European city.
a. For such a business, How may a geodemographic system be used for marketing decision making?
b. For such a business, How may a geodemographic system aid marketing research design?
c. Present the case of which of the above systems would best suit such a business.
2. Call in at a supermarket or store that operates a reward or loyalty card scheme that requires you to apply for membership. Pick up an application form and examine the nature of questions you are expected to answer.
a. What marketing research use can be made of the data collected from this application form?
b. Evaluate the design of this form and make recommendations on how the nature of questions could be improved.
3. You are a marketing manager for Carlsberg beer. One of your major customers is a supermarket that uses a loyalty card scheme to electronically observe its customers.
a. What would this supermarket know about beer buying behaviour through their scheme?
b. If they would not share this with you, evaluate any marketing research techniques that you think could generate the same knowledge.
4. Visit the SPSS website www.spss.com and follow a link to their ‘Predictive Analytics’ products. Write a report on how marketing research may feed into and/or feed from predictive analytics in either a bank or major retailer.
5. In a small group discuss the following issues: “what ethical problems exist with the use of marketing databases for marketing researchers?” and “if on ethical grounds, a marketing researcher refused to utilise the benefits of marketing databases, what inherent weaknesses may exist in their research designs?”

Figure 5.1 Methods of segmenting markets
Stage 1: Existing customer database
¯
Stage 2: Geodemographic profiles
¯
Stage 3: Combine customer data with geodemographic profiles
¯
Stage 4: Add survey data
¯
Stage 5: Create `own’ consumer profiles
¯
Stage 6: Use of the datawarehouse
Figure 5.2 Stages of development in using databases and survey data to build profiles of consumers and model marketing decisions
Notes
[1] The countries where Experian are able to access data and build their groups
[i] McElhatton, N., Customer insight in stereo Research World, (December 2004), 21
[ii] McElhatton, N., Customer insight in stereo Research World, (December 2004), 20
[iii] Evans, M., Nancarrow, C., Tapp, A. and Stone, M., ‘Future marketers: future curriculum; future shock? Journal of Marketing Management, 18 5-6, (2002) 579-596
[iv] Dawe, A., `Integration is the thing’, The Times – Special Edition e-CRM (11 April 2001), 2–3.
[v] Sumner-Smith, D., `Getting to know clients lifts profits’, The Sunday Times (26 September 1999), 17.
[vi] Fletcher, K., `The evolution and use of information technology in marketing’, in Baker, M.J. (ed.), The Marketing Book, 3rd edn (Oxford: Butterworth-Heinemann, 1994), 352.
[vii] Mistry, B., ‘A question of loyalty’, Marketing, (2nd March 2005), 42
[viii] Havermans, J., Do-it-yourself surveys offer opportunity for MR, Research World, June 2004, 20-21
[ix] Kilby, N., and Bedwell, R., Women stand by their stores, Marketing Week, June 2nd 2005, 32-33
[x] Lawson, J., Cruising for customers’, Database Marketing, (December 2003), 29-30
[xi] An example of fusing research data with databases is presented in Leventhal, B., `An approach to fusing market research with database marketing’, Journal of the Market Research Society 39(4) (October 1997), 545–61.
[xii] Sleight, P., Newest neighbourhoods, Database Marketing, (June 2005) 31-34
[xiii] Man, D., `You say “warehouse”, I say “database” …’, Bank Marketing 29(4) (April 1997), 37.
[xiv] Man, D., `The faqs on datawarehousing’, Bank Marketing 29(4) (April 1997), 38.
[xv] Lawson, J., No rocket scientists required, Database Marketing, June 2003, 30
[xvi] Jambu, M., `Data mining for better knowledge’, A case study for the telecommunications industry, ESOMAR, Power of Knowledge Conference, Berlin (September 1998).
[xvii] McElhatton, N., `Raiding the data bank’, Research (September 1999) 28–30. See also the arguments in Brand, C. and Jarvis, S., `Mind games – the new psychology of research’, Market Research Society Conference, 2000.
[xviii] Pinnell, J., Customer relationships – Manage, measure or just understand? ESOMAR (Jan 2003) www.warc.com/print/78475p.asp
[xix] McElhatton, N., Customer insight in stereo Research World, (December 2004), 21
[xx] Hodgson, P., `Databases: the time for decisions is nigh’, Research (October 1993), 19.
[xxi] ‘Loyalty for sale’, Financial Times (18 September 1998).
[xxii] Smith, D.V.L. Bridging the gap between data and decision, Research World, (November 2004) 20-21 and Savage, M., `Downstream danger’, Research (May 2000), 25–7.
[xxiii] For illustrations of neural networks and decision trees in marketing see: Garver, M.S., Using data mining for customer satisfaction research, Marketing Research, (Spring 2002) 8-12