1. Determining the attributes to measure
No matter which scale is used the crucial factor to get right is the wording of the items against which the attitude is to be measured. As with all questionnaire research, if the item is not measured it cannot be analysed, and if important attributes are not included then the analysis could be totally misleading.
If there is no existing set of attitude or attribute dimensions that have been proven to represent the issues in the market under consideration, then they will need to be developed.
Ideally the dimensions should be developed through a preliminary stage of qualitative research, designed specifically to determine the range of emotions, attitudes and perceptions that exist and that are relevant to the study and its objectives. The principal purpose of the preliminary study is to provide the attitude dimensions that are to be measured for strength of agreement in the quantitative survey. This stage can also be used to develop some preliminary hypotheses about attitudinal segments that might exist in the market, which the quantitative survey can then test.
If it is not possible to carry out a preliminary stage, the dimensions must be collated from elsewhere. Previous studies in the same area are the best place to start, even if they were not designed to meet precisely the same objectives. Any similar work carried out previously by the client should be examined.
Sometimes, though, it comes down to experience and brainstorming, in an effort to try to generate every possible attitude, emotion or image perception that might exist and might need to be included in the questionnaire. This approach has obvious dangers:
- New attitudes that have not yet been identified may be omitted, which will tend to lead to a continuation of the existing perceptions of the market, rather than providing new insight.
- Something important may be overlooked completely.
- The wording used may not be that used by the respondents.
- In the absence of any information as to what is and what is not important, there will be a tendency to produce too many dimensions in an attempt to ensure that everything is covered.
To counter this last point it is not unusual for a preliminary survey to be conducted that concentrates principally on the large set of attitude dimensions that have been initially generated. Most other questions are omitted from this questionnaire in order to make it manageable for the respondents. However, care must be taken not to alter the context of the attitude question by omitting preceding questions such as those about the respondent’s behaviour in relation to the topic. Techniques such as principal component or factor analysis are then used to reduce a large battery of attitude dimensions to a smaller, more manageable set that can be included in the questionnaire. There is a danger here, though, that small differences in attitude dimensions – ones that were specifically introduced in the brainstorming because they are important – get excluded because the purpose of the factor analysis is to produce broader, underlying attitude dimensions. It is important, therefore, to follow any reduction process by a further review of the dimensions and reinstate those of particular importance, or showing particular nuances of difference, that have been removed.
There exist sources such as the Handbook of Marketing Scales (Bearden and Netermeyer, 1999) that provide lists of dimensions for a range of different attitudinal subject areas that have been used in published studies. They are a useful starting point for someone compiling an attitude battery, or can be used when looking for standardized wording or checking that the compiler has not overlooked an important dimension. Before adopting a complete set of standardized scaling dimensions, however, users should ensure that they cover all aspects of the topic under consideration in their study.
2. Number of attributes
If the number of statements exceeds the respondents’ boredom threshold, the likelihood of pattern responding is increased. The threshold will vary according to people’s interest in the subject.
The size of the statement battery is something that the researcher should consider carefully. Clearly there must be a sufficient number of statements to address adequately all of the attitudes under consideration. If possible, there should be several statements for each attitudinal dimension to enable the researcher to cross-check responses for consistency within respondents. The number of statements before fatigue sets in will vary according to the level of interest of the respondent in the subject. However, the maximum number in one battery is rarely more than about 30 before a respondent’s attention begins to wander. If questionnaire writers are unsure, they should ask themselves whether they could themselves maintain concentration throughout a battery of 200 statements about, say, greetings cards.
If, despite all attempts to reduce the number of statements, it is not possible to cover the required attitudinal dimensions without producing a formidable battery of statements, it can sometimes be possible to split the statements into two batteries that are located at different points in the questionnaire. The statements should be split so that the two batteries cover different sets of underlying attitudinal dimensions, and if possible this should be explained in the introduction to the question. Without this precaution, there is a danger that respondents will believe when they are presented with the second battery that they are being asked the same questions again and will not take sufficient care.
Nevertheless, with a battery of statements of any size it is inevitable that some respondent fatigue will set in. Statements at the beginning of the battery will be given more careful consideration than those towards the end. The dangers of this type of response order bias and how to deal with it are discussed in Chapter 7.
3. Indirect techniques
The difficulty that people have in recognizing – let alone accurately articulating – their emotions and feelings about brands has led to a number of techniques that approach the issue indirectly. For example, instead of asking respondents to associate image dimensions with brands, techniques have been established that associate the brand with picture stimuli, which in turn are established as having certain emotional associations. The respondents feeling towards the brand can then be evaluated, even if the respondents do not consciously recognize those feelings themselves.
Most of the techniques of this type, however, are proprietary and have a specified set of questions. They are therefore outside the scope of this book.
4. Pictorial techniques
Many of the indirect techniques use pictorial stimuli either to convey a personality type or emotion with which respondents are asked to associate needs or brands, or to help the respondents to express how they feel in a way that would be difficult for them to do verbally.
The latter type of response has been used principally to evaluate people’s emotional response to advertising. Respondents are shown a series of depictions of emotions and asked which best represents how they felt as they watched the advertisement. This type of approach relies on a theoretical framework that encompasses the full range of emotions, and that defines the emotions to be depicted.
The use of pictorial stimuli is fraught with difficulties. It is often the case that variations between pictures, other than those intended, bias the responses as people react to these differences rather than those the questionnaire writer wishes them to. Figure 6.4 shows the solution to this adopted by NFO Worldgroup, which uses a series of collages of people of different age, gender and race to depict each of these emotions. A more recent approach to this issue (Wood, 2007) has been the use of a single androgynous face used in a series of pictures to portray a range of emotions, avoiding this type of bias.
Source: Brace Ian (2018), Questionnaire Design: How to Plan, Structure and Write Survey Material for Effective Market Research, Kogan Page; 4th edition.