Outline of interview structure

The main purpose of a cognitive research interview is to discover the knowledge content of a person's mind. The interview itself should attempt not to change existing knowledge but only to retrieve it. Inferences in recall should be kept to a minimum. At the same time the problem of forgetting has to be dealt with; and here, the positive and negative effects of part listings have to be traded off. As a consequence of these considerations, some structure has to be imposed on the interview procedure (Patton 1990:43, 283, 288, 376, Cohen 1989:6-7, Sampson 1986:33, Merton and Kendall 1945-6:545-7).

This study applied a question guide that divided the interview into four parts (see Appendix 3.1): in the first part of the interview, only open questions were asked and no part listings were used; in the second and third parts of the interview, part listings were used; the fourth and final part concluded the interview with a self-rating task and offered interviewees the opportunity to comment on further issues they wished to raise.

The first part of the interview contained only open questions drawing on free retrospective recall. No specific items, either products or product attributes, were singled out for an interviewee to comment on. No part listings were used. Thus, the recall restraining effect of part listings was avoided, but the problem of forgetting was not dealt with. The retrieval of knowledge content on green shopping was organized around products rather than product attributes. For supermarket products, product-related cognitive processing rather than attribute-related processing has been found to be dominant; examples of products can be expected to be easily recalled by an interviewee (Park et al. 1994: 79, Alba et al. 1991:7-10, 21-5, Lynch and Srull 1991:111, 122, Rothschild 1987:569).6 Hence, as a first step, green product examples were collected from an interviewee without the interviewer probing more deeply as to why and how an interviewee had come to view a product as green (in terms of product attributes). These examples of green products were then discussed with the interviewee who was asked about his or her subjective understanding of greenness. That kind of phased approach, starting with the recall of product examples that then provide the basis for the interviewee to discuss how he/she perceived green product attributes, appears to restrain inferences in recall best (Alba and Hutchinson 1987:433, Smith and Houston 1986:505, Markus 1977:65).

In the second part of the interview, a part listing of products was used. Problems related to forgetting were reduced by this. The purpose of a part listing is to provide retrieval cues. Part 2 (and also part 3) of the interview required the interviewee to perform a recognition task rather than a free recall task. Comprehensive product part listings are likely to contain more than forty product categories (see Figure 3.1). Both for reasons relating to the efficient collection and analysis of data and for reasons relating to the span of attention an interviewee is able and prepared to offer for the purpose of an interview —the interview was expected to last about 30 minutes—no all-out testing of product categories was possible. Only those product categories were included in the part listing that had frequently been mentioned as examples of green products in the pilot studies or in the literature on environmentally oriented consumption (Worcester 1994a:5, Mintel 1992:2, CA 1989:433). In the end, the part listing comprised eleven products. Such selective lists of product categories are commonly applied in consumer behaviour research (Hoyer 1986: 33-5). The part listing was applied by default, i.e. there was prompting for only those products of the part listing that had not already been discussed in the first part of the interview. Once an interviewee had responded positively to a product from the prompting list, a follow-up questioning comparable to that in part 1 was conducted.

In part 3 of the interview, a part listing was applied that comprised different types of information cues a consumer might have drawn upon to assess the greenness of a product. The part listing was constructed on the basis of the findings of the pilot studies and a literature review on consumer behaviour (Billig 1994:1619, 180-4). That kind of enquiry was put at the end of the interview because of its higher potential to cause inferences in recall. A general discussion of information related to the perceived greenness of a product is more prone to lead to inferences in recall.

In a final, fourth part, the interview was concluded with a self-rating task. The interviewee was asked for an overall self-assessment regarding the greenness of his or her shopping. On a scale ranging from zero to 100, interviewees had to rank themselves with regard to actual green consumption behaviour (see Figure 3.2). They were then given the opportunity to comment on any further issue they wanted to raise in relation to environmentally friendly consumption.

In addition to green consumers, a group of 'non-green' consumers was interviewed. Basically, the same interview procedure was applied, but with the omission of parts 1 and 4. Part 1 of the interview procedure, the free recall of green products and the discussion of them, had necessarily to be left out, as had a self-rating task regarding the greenness of the consumers' shopping. Parts 2 and 3 of the interview procedure used the same part listings, but (in a hypothetical fashion) with regard to intended behaviour. There were different start-ups to the interview with non-green consumers depending on how they had been recruited, either on the doorstep or through recommendations by previous interviewees.

The interviewer tried to minimize an inferences in recall bias by structuring the interview in a fairly concrete fashion with the focus of enquiry on product examples. If inferences in recall occurred despite these attempts to minimize them, the interviewer should be able to recognize them. For the recognition of inferences in recall that transcend mere constructive recall, a number of criteria have been suggested, such as

Food products

Dairy products:

milk yoghurt cream cheese butter etc.

Plant products: vegetables, fruit pulses, grains: rice, lentils coffee, tea etc.


meat, sausages etc.


drinks: mineral water, beer, wine, soft drinks, orange juice etc.

Non-food products


hair spray, deodorant: sprays shampoo, shower gel cosmetics: cream, lotions, cleansers soap, hand cleaner toothpaste etc.

Cleaning products/detergents: washing powder/laundry detergent fabric softener washing-up liquid cleaners: floor cleaner, toilet cleaner/bleach, window cleaner, surface cleaner polish etc.

Paper products: toilet paper tissues: handkerchiefs kitchen paper nappies women's sanitary products etc.

Bakery products: bread biscuits: cakes flour etc.

Others: bin bags etc.

Others: sweets spreads: jam, margarine eggs baby food ready made food etc.

Figure 3.1 Products consumed on a daily basis long periods of silence, the absence of slang, or the absence of redundancy, none of which is common in ordinary speech (Bettman 1979b:232; also Markus 1977:65). The researcher's impressions with regard to these issues provided the basis for assessing whether inferences in recall had actually occurred. Also, interviews with non-green interviewees on intended green shopping behaviour provided post hoc an idea of what extreme 'green' rationalizing and hypothesizing looked like. Since the non-green interviewees had very low levels of actual green shopping experience, they had to make shrewd guesses, i.e. invent hypotheses and rationalizations, on what makes a product green. (Chapter 5 returns to this issue of identifying 'rationalizing' when findings on the non-green interviewees are compared to the findings on

I take environmental considerations into account when I shop for groceries

Figure 3.2 Self-rating scale green interviewees; see especially the section on schematic knowledge, where stereotyping by non-green interviewees is discussed.) In general, it appeared that inferences in recall presented an insignificant problem for the research conducted.

Continue reading here: Sampling of interviewees

Was this article helpful?

0 0