Wilson’s application of Appraisal Theory to market research interviews
As implied above, many commercial research methods begin with a qualitative interview: individuals, pairs or groups speak to a facilitator who records (either electronically or in note form) what they say. While it is possible simply to accept statements of attitude elicited during the course of a qualitative interview, a trained linguist can gain more access to the depth and strength of people’s feelings through understanding the unconscious choices in their language. Wilson (2011) has shown how the analytical framework offered by Systemic Functional Grammar (e.g., Martin 2003; Martin and White 2005) and Appraisal Theory (Hunston and Thompson 2003) can facilitate this. He developed a detailed framework for ranking the strength of feeling behind utterances in a focus-group setting by looking at evaluative language and its place in conversation. His subject matter was market-research groups conducted by a large food and household-product company, and, in particular, sessions in which participants were asked to discuss their preferences for various container shapes for men’s anti-perspirant. While this may seem limited in scope, the framework itself is widely applicable to qualitative research groups in general, from product and communication evaluation to political attitude.
Wilson (2011: 96) suggests the following categorisations of attitude:
- - Affect: how a speaker is emotionally disposed to the subject of the communication (e.g., I definitely like that one the best);
- - Judgement: how the subject of the communication compares to accepted norms and values (e.g., It’s something out of the ordinary, isn’t it?);
- - Appreciation: how the subject of the communication creates an impact on the speaker in terms of form, appearance and aesthetics (e.g., The narrower neck makes it more feminine).
However, an important point of Wilson’s research is that not all utterances come complete with some kind of obvious (i.e., lexically marked) evaluation. In most, the evaluative content is much less overt. For example, the following example could be a case of appreciation (“impact in terms of form, appearance and aesthetics”), but it is not easy to see whether the force is positive or negative.
- It reminds me of those little perfume bottles that you get with the squeezy bit on it.
Here, it is necessary to look at the surrounding context, where the speaker says:
- So probably more appealing because it looks a bit more old-fashioned and that’s what’s fashionable at the moment.
The framework also allows for strengthening or weakening of the evaluation, indicating what Wilson terms degree of engagement. Here, we show a weakened evaluation followed by a strengthened one:
- - I don’t know, it just doesn’t appeal to me really.
- - That’s definitely a very feminine shape.
As well as examining the gravity and status of evaluations of a wide variety of kinds, the framework goes further to expose how far an evaluation’s context strengthens or weakens it as an indication of speaker attitude. This is because Wilson also takes account of an evaluation’s role in conversational structures by using research on turn-taking and preferences for the forms that turns will take (e.g., Pomerantz 1984; Brown and Levinson’s 1987 Politeness Theory, and the work of many researchers following them). His framework thus acknowledges that utterances are not made in a void: instead, they are heavily conditioned by the need to maintain agreement between speakers as far as possible, to follow predicted turn sequences, and to avoid threatening the other speaker’s (or the interviewer’s) “face”. This leads Wilson (2011: 66) to suggest that “more importance can be placed on some evaluations than others, based on the difficulty or cost associated with making that evaluation [sic]”. Thus, for example, the interview might contain a turn such as the following between speaker A and B (discussing bottle shape):
- - Oh I really don’t, no, I don’t like that one at all, it’s far too bulky.
- - Yeah I see what you mean, but I was going to say that I quite like the style of the erm the grips on that.
Although speaker B is hedging their evaluation (“Yeah”, “I was going to say”, “I quite like”), in terms of Wilson’s framework both speakers are making relatively strong evaluations. Speaker A is informing speaker B of their opinion, possibly in response to the interviewer’s prompt, but the opinion is “out of the blue” in that it is not predictable from any previous utterance. Speaker B, according to Pomerantz (1984), is strongly obliged both to respond and to agree. However, while the hedging and the first part of the turn indicate that they clearly understand that obligation, they then offer liking for the “style of the grips” rather than the expected agreement, which is therefore not preferred behaviour. That means speaker B’s evaluation, too, is a strong one and a reliable indicator of their feelings.
Because it allows for the difference between completely spontaneous evaluations and those prompted by previous context, the application of turn-taking and conversation structure analysis controls for two issues in focus group or interview situations. These are, first, the impact of the moderator’s questions on subsequent turns and participants’ willingness to take the floor (Myers 2007: 81), and, second, the relationship between more and less dominant speakers in a group.
However, as Myers (2004) points out, an obvious drawback of applying a framework such as Wilson advocates is that in realistic market research there is often little understanding of the need to fully transcribe interview or focus group data or to analyse it in detail - even if there were time to do so. This means that a great deal of the detail upon which the framework relies is lost unless the budget and timeframe allows for close transcription. However, Wilson’s research does indicate that, given sufficient care, focus groups and interviews can yield richer insights through linguistic analysis - which might perhaps support their reputation as a research approach.