A Market of Opinions: The Political Epistemology of Focus Groups

Introduction

Provoking a conversation among a small group of people gathered in a room has become a widespread way of generating useful knowledge. ¹ The focus group is today a pervasive technology of social investigation, a versatile experimental setting where a multitude of ostensibly heterogeneous issues, from politics to economics, from voting to spending, can be productively addressed. ² Marketing is the field in which the focus group has acquired its most visible and standardized form, as an instrument to probe and foretell economic behaviour by anticipating the encounter of consumers and products in the marketplace. ³ But whether they are used to anticipate consumer behaviour in a laboratory-like setting, or to produce descriptions of political attitudes, conversations elicited in the ‘white room’ of the focus group are relevant to a striking range of objects of social-scientific inquiry. ⁴

The observation of contrived groupings of research subjects in ‘captive settings’ is of course a familiar source of knowledge in the social sciences, but there is something peculiar to the focus group as a research technology. In focus groups, knowledge is generated in the form of opinions. Moreover, a group dynamic is used to bring into existence a series of relevant individual opinions; the peculiar form of social liveliness of the focus group is meant to ‘produce data and insights that would be less accessible without the interaction found in a group’ (Morgan, 1988: 12). Both the productive qualities and methodological quandaries of the focus group originate in its special form of liveliness. The peculiar politics and epistemology of a focus group conversation derive from the tension implied in using a group to engender authentically individual opinions. Moderators are in charge of resolving this tension: they must make the conversation conducive to the expression of private and idiosyncratic views, while preventing the focus group from rising to the status of a ‘collective;’ they are called to structure a process of interaction conducive to the elicitation and elucidation of the most private of views, while reducing to a minimum the residuum of ‘socialness’ left over from the process. As a professional group moderator describes it:

Researchers hope to externalize the silent ‘running commentary’ of consumers by means of an intently managed group discussion, to translate a series of inaudible monologues into a visible conversation. They provoke an exchange so as to bring to light the inner qualities of consumers.

Knowledge about people is extracted from the opinions elicited from them – opinions that are freely expressed by the subjects, yet structurally incited by the setting. ⁵ Those opinions are then selected, categorized and interpreted by the focus group researcher and fed into production and marketing strategies. ‘Illustrative opinions’ are filtered from the wealth of talk generated in the discussion, to be quoted verbatim or paraphrased in the research reports circulated to clients and other relevant audiences. Thus, opinions generated in the ‘white room’ are read, interpreted, and discussed by managers and marketers who were not present in the original conversation and are in no position to directly assess their authenticity or relevance. The statements produced in the unique environment of the focus group enter a long chain of quoting and rephrasing, and reverberate into other actors’ market strategies. The ultimate product of a focus group conversation is a series of tradable opinions – statements that are generated in an experimental setting but can be disseminated beyond their site of production. Opinions elicited from focus group participants thus help constitute particular marketplaces.

Producing opinions of such value and mobility is a highly complex technical process. A focus group can generate a multitude of objects that, while seemingly identical to relevant opinions, are in fact radically different kinds: false opinions, induced judgments, or insincere beliefs, all of which appear profusely in the course of a focus group discussion – especially in a poorly run one. These deceptive statements must be sorted out and expunged so as not to lead researchers and their audiences astray. The task of the moderator is to manage the focus group discussion so as to limit the proliferation of irrelevant or inauthentic viewpoints; to foreground tradable opinions against the background noise that is inevitably generated in the experimental situation.

The purpose of this chapter is to draw attention to some of the strategies utilized by focus group moderators to carry out this task of extracting tradable opinions out of experimentally generated conversations. In so doing, we can regain a proper appreciation of the extent to which categories such as ‘relevant opinion’ or ‘consumer preference’ are problematic – and not simply or primarily to the external observer, but to the actors who are professionally trained to elicit and recognize them, the focus group moderators.

My account will be limited in a number of important ways. The manufacture of opinions in a focus group starts with the assembling of a group of adequate research subjects and a meeting with one or more moderators, but the ‘focus group chain’ comprises a long sequence of exchanges and analyses beyond this initial encounter. This chapter, however, will only investigate the initial experimental moment, when research subjects and moderators come together in the physical setting of the focus group ‘white room.’ Moreover, I will analyse this encounter solely from the perspective of the moderators: my analysis is based on the moderators’ own technical literature – the training manuals, methods handbooks, autobiographical accounts, and other documents in which they lay out their own philosophy of ‘good practice’ and a portrayal of the ‘good moderator.’ I do not attempt to examine the focus group discussion from the point of view of the research subjects, nor will I draw extensively on analyses of the patterns of interaction between subjects and moderators that actually emerge in a focus group, a dimension of the focus group encounter that others have studied at some length (Myers, 1998 and 2004; Myers and Macnaghten, 1999; Puchta and Potter, 1999 and 2003). The chapter is thus limited to descriptions of the craft of moderation that professional moderators have put into writing. ⁶ Through this literature, I try to reconstruct an ideal moral epistemology of moderation. In particular, I try to capture the political constitution of an experimental setting in which individual attitudes are elicited and market behaviour is routinely anticipated.

The chapter is organized around three themes, all of them topics that social scientists have frequently raised in relation to the production of scientific knowledge under experimental conditions: 1) the distinction and balance between naturalness and artificiality in the focus group setting, and the embodiment of this distinction in the moderator's skills and abilities (or, rather, in the accounts that moderators give of their own craft); 2) the co-production of knowledge and particular forms of social order, or the political constitution of the focus group – a constitution that ideally, I will argue, takes the form of an isegoric assembly; and finally, 3) the role of material artifacts and the physical arrangement of the setting in the organization of the ‘focus group chain’ as a technology of knowledge production. The chapter concludes with a call to make the production of opinions a proper object of sociological investigation, in the same way that the creation and circulation of knowledge has long occupied a central place in the agenda of sociological research.

Chameleonic moderators

A fundamental dichotomy runs through the technical literature on focus group ‘facilitation’ and serves as an object of interminable reflection for moderators: should the focus group be seen as a natural phenomenon, closely resembling a naturally occurring (or ‘casual’) conversation, or should it rather be treated as the highly artificial outcome of an experimental intervention? Moderators’ opinions on this issue obviously vary. Some emphasize the similarity between the kind of conversation they hope to encourage, and naturally occurring talk. ‘The moderator,’ one argues, ‘should be viewed by the group as the authority figure in the room but also as the type of person with whom they would like to have a casual conversation’ (Greenbaum, 1988: 51). ‘The frequent goal of focus groups,’ another stresses, ‘is to conduct a group discussion that resembles a lively conversation among friends or neighbours, and most of the problems come from topics that fail to meet that goal’ (Morgan, 1988: 22). The productive liveliness of the interaction depends, according to this view, on a successful replication of the conversational patterns of friends or neighbours. Yet, as Agar and MacDonald point out, ‘it is not automatic that a group of strangers will have a ‘lively conversation’ about anything’ (Agar and MacDonald, 1995: 78), and in fact many moderators choose to emphasize the seemingly opposite view: that whatever takes place in a focus group should be understood and analysed as the result of a radical experimental intervention, that the focus group is a highly artificial product – from which, however, valid knowledge about the natural world can be extracted. A historian of the focus group form puts it as follows:

Judging from the body of literature analyzed for this article, the view of the focus group as an artificial encounter is more widely held among practitioners (or, rather, among the authors of methodological texts), but the recognition of artificiality is often accompanied by caveats and justifications that suggest a structural orientation towards the model of a casual, natural conversation as the regulatory ideal for their practice. Moderators can and will explain the departure from naturalness as inevitable and useful, but in addressing the differences vis-à-vis naturally occurring talk they highlight the relevance of the latter as the ideal of talk to which they aspire. In other words, while most moderators will readily acknowledge that the focus group is not a natural setting, very few will be willing to give up the claim to conduct a ‘naturalistic inquiry’ (Morgan and Krueger, 1993: 8). ⁷ One could argue that focus group moderation is to them akin to an Aristotelian ‘perfective art’ – an intervention that, by removing the obstacles in their way, ‘perfects natural processes and brings them to a state of completion not found in nature itself’ (Newman, 2005: 17).

Whether they choose to emphasize the naturalness or the artificiality of the focus group in the descriptions of their trade, most moderators would nevertheless argue that, in the practical conduct of a focus group, both dimensions need to be tackled and made compatible; that the strength of the focus group as an engine for the production of tradable opinions rests precisely on the ability to combine these apparently contradictory dimensions; and that the point at which these trends are unified is in the very figure of the moderator. This is paradoxical, since, as was pointed out in the quote above, it is the very presence of the moderator that renders the situation hopelessly artificial. Yet, in their methodological writings, particularly in the portrayals of the ‘good moderator’ put forward in the technical literature, one finds a pervasive effort to combine in the kaleidoscopic identity of the moderator the apparently irreconcilable demands of naturalness and artificiality. This combination of conflicting qualities surfaces in long, colourful descriptions of the moderator's persona:

In these characterizations, the moderator embodies – in his skills but also through his personality – the conflicting burdens of objective detachment from, and natural empathy with the research subjects. Moderating is always more than a science – or a set of easily formalizable techniques:

The prevalence of artistic metaphors is worth noting here and in other instances:

It is also, and crucially, an art mastered only by the few:

The tension between the natural validity of the product the focus group strives to generate and the experimental nature of its extraction is thus translated into multidimensional portrayals of the ‘good moderator,’ of the skills and personality traits best attuned to this research practice. The tension is not resolved, but rather given a new form in the self-discipline of a moderator simultaneously capable of both attachment and detachment, of leading a natural conversation and generating experimental observations at the same time.

The emphasis in the technical literature on the personality of the moderator, in addition to his technical skills, as the key to a successful focus group is a relatively new phenomenon, likely connected to the professionalization of focus group research and the effort to market the value added by moderators. In the descriptions of focus group moderation offered by the pioneers of the trade one finds a very different tone. In their seminal 1946 article on ‘focussed interviews’ (the term they coined for the technique), Robert K. Merton and Patricia Kendall placed the emphasis on the recurrence in the focus group situation of specific patterns of interaction, and on the fundamental teachability and transferability of the techniques necessary to manage these typical situations:

One detects in their descriptions a view of the moderator as a technician of sociological investigation. ‘We have found,’ Merton and Kendall continue, ‘that the proficiency of all interviewers, even the less skilful, can be considerably heightened by training them to recognize type situations and to draw upon an array of flexible, though standardized, procedures for dealing with these situations’ (Merton and Kendall, 1946: 545). Hence the term ‘interviewer,’ which Merton, Kendall and other members of the Bureau of Applied Social Research used throughout their writings to refer to the moderator of a discussion. ‘Interviewer’ emphasizes an essential continuity between different research interventions – from asking standardized questions drawn from a questionnaire, to the managing of a group of research subjects. The ‘focussed interview’ could be applied to individuals as well as to ‘groupings,’ and it always involved a recurring question-and-answer pattern of interaction. ⁹ The conception of the moderator as interviewer represents the polar opposite of the open-ended, idiosyncratic practices denoted by names such as ‘facilitator,’ or ‘qualitative research consultant.’ To the social-scientific pioneers of the group discussion, the interviewer is above all a trainable instrument of social research, rather than the chameleon-like master of an arcane art of conversation. ¹⁰

Orienting the subjects’ reflexivity

If moderators incorporate in the descriptions of their craft an array of diverse traits and features, including a strong element of self-discipline, they devote most of their technical literature to the management and disciplining of others: their research subjects. The self-discipline of the moderator and the disciplining of subjects are but the two sides of the same coin. In both cases the question is, once again, how to strike an adequate balance between naturalness and artificiality. In the case of the research subjects, the key to this balance is the proper management – the incitement, orientation and taming – of their reflexivity.

In short, the fundamental problem for moderators is how to turn the research subjects away from the experimental features of the setting. That is, how to prevent subjects from explicitly attending to their peculiar context, the focus group ‘white room,’ so that their actions – in this case, the opinions they express – are not direct responses to their being in that particular setting and can be projected beyond their site of production. Moderators have to steer the participants’ inevitable contextualization of their own situation and statements (Myers, 2004: 56–66), and neutralize the attention they are bound to pay to the experimental nature of their grouping.

The technical literature addresses this general problem through a series of ‘type situations’ to be tackled by the well-trained moderator. Perhaps the most obvious example of a participant's orientation towards the experimental setting is the problem of sabotage. Sabotage is a reflexive response, and a highly disruptive one at that. It occurs when, for whatever reason (because they dislike the research question, or the way the moderator goes about extracting their opinions, or simply their being there) participants actively try to upset the proceedings of the focus group. Dissatisfaction with the experimental situation can be expressed in a multitude of ways. In some cases, research subjects avail themselves of the mechanisms of the focus group to generate a disruptive outcome – for instance, by expressing what, from the point of view of the moderator, will turn out to be a series of useless or irrelevant opinions. ¹¹ More frequently, sabotage is in the moderators’ literature equated with a lack of respect for the basic etiquette of civilized conversation. Bullying other participants, showing excessive aggressiveness in stating one's own views, or trying to monopolize the conversation are all ways of contravening the rules under which the focus group is conducted. ‘There are some people who just are unpleasant,’ a moderator remarks. ‘They must be removed if they become intimidating to other members of the group’ (Morrison, 1998: 211). The techniques used to deal with this kind of behaviour run the gamut from withdrawing eye contact, in the hope of silencing the offending participant, to directly removing the disrupting individual, or even calling off the meeting.

Unproductive or disruptive reflexivity is sometimes not a matter of participants behaving badly, but of the researchers enrolling the wrong kind of participant. This is the case with the ‘professional respondents’: people who either make a living out of participating in focus groups, or simply enjoy them too much and manage to participate in too many. They represent a serious problem because they are too familiar with the setting. While detecting and weeding out these ‘repeaters’ is nowadays largely a matter of computerized screening of research subjects, the moderator must always be on the look-out for signs of excessive familiarity or comfort on the part of the research subjects. Research subjects that appear too relaxed or gregarious are to be treated with suspicion. ‘Sometimes a focus group feels fishy,’ a moderator notes:

The repeater's familiarity with the focus group setting generates a sense of ease and naturalness that is at odds with the ability to extract tradable opinions from him. Familiarity is, like the conscious sabotage, a form of orientation towards the specific features of the focus group setting – an orientation characterized in this case by fluency and skill rather than by awkwardness or aggressiveness, but equally detrimental to a proper balance between artificiality and naturalness. ¹²

A third situation, more subtle and widespread, is not related to the kinds of individuals who take part in the discussion, but to the kinds of opinions they express. Merton and Kendall described this in their 1946 article as the problem of direction: the process by which the actions of the moderator end up ‘modifying the informant's own expression of feelings.’ When this occurs, ‘[t]he interview is no longer an informal listening-post or ‘clinic’ or ‘laboratory’ in which subjects spontaneously talk about a set of experiences, but it becomes, instead, a debating society or an authoritarian arena in which the interviewer defines the situation’ (Merton and Kendall, 1946: 547).

For Merton and Kendall, ‘directed’ opinions were those not spontaneously expressed in the research subject's own ‘frame of reference.’ This is why moderators must be careful to use ‘nondirective techniques’ and to give the research subject ‘an opportunity to express himself about matters of central significance to him rather than those presumed to be important by the interviewer’ (Merton and Kendall, 1946: 545). It is the possibility of generating nondirective outcomes that makes focus groups a more adequate technology to elicit authentic opinions than polling and other highly formalized research tools. This is a central tenet of the moderators’ epistemology, going back to the foundational uses of group interviews by Merton and others: ‘in contrast to the polling approach, [the focussed interview] uncovers what is on the subject's mind rather than his opinion of what is on the interviewer's mind, and ‘it permits subject's responses to be placed in their proper context rather than forced into a framework which the interviewer considers appropriate’ (Merton and Kendall, 1946: 545).

The moderators’ technical literature also discusses the issue of ‘direction’ as the problem of ‘moderator demand,’ defining it along similar lines to the ‘experimenter demand’ in experimental psychology: situations ‘where the respondent or subject guesses what is going on, realizing what is expected of him or her and ‘helping’ the experiment along by performing in ways that they think the experimenter would like’ (Morrison, 1998: 182). Most participants in a focus group are certainly going to realize that something is going on, if not what exactly is going on. As Myers argues, a focus group is in fact ‘two hours of “What is it that's going on here?“’ (Myers, 2004: 56). The focus group meeting is a highly contrived occasion, and research subjects are obviously aware that the fundamental purpose of their being there is to get them to talk about matters of relevance to the researcher. More often than not, they will try to be ‘helpful.’ While it might seem the polar opposite of sabotage, helpfulness is, from the point of view of moderators, a similarly disruptive orientation towards the experimental nature of the encounter. As Morrison writes:

This is why, rather than being a negative quality, non-direction needs to be actively engineered into the behaviour of the moderator and into the responses he elicits from the research subjects. Moderators need to forestall the natural propensity to help. Research subjects will always try to anticipate what is expected and desired from them, and are very likely to act (ie, express views) on the basis of these assumptions. The task of the moderator is thus much more complicated than simply letting the conversation run its ‘natural course,’ or leaving respondents to their own devices. He or she has to manage the expectations of participants carefully so as to generate nondirective outcomes that are conducive to the research. The role of the moderator is thus to provoke preconceptions that would suit her research purposes, without allowing research subjects to be conscious of their own helpfulness. The moderator must, through self-discipline and the careful management of information, elicit what Merton described as the ‘self-betrayals’ and ‘self-revelations’ of the research subjects. He must tame and channel the subjects’ unavoidable awareness of the research situation towards goals beyond their grasp.

Influence: the moderator as political philosopher

In his study of jury deliberations, Harold Garfinkel analyzed the rules of judgment that jurors were expected to use in producing a legal decision (Garfinkel, 1984). He discovered that, while jurors simultaneously entertain everyday and official rules of social judgment, they use the ‘official line’ to produce retrospective accounts of the decisions they arrive at, and of the process by which they arrive at them. The focus group shares important similarities with the jury setting – another experimental setting designed to generate relevant opinions – but it presents even more telling differences. Focus group participants are seldom asked to justify their opinions on the basis of a set of official rules. They will normally be asked by the moderator to elaborate upon and develop their opinions, to extend their contributions by pursuing the ramifications of their initial statements, but this is quite different from being asked to produce accounts of their opinions in accordance with a set of formalized procedures. A second, even more fundamental difference between juries and focus groups is that, while jurors are asked to abandon their everyday patterns of judgment in reaching legally relevant opinions, the participants of a focus group are expected to come into the experimental situation with all their preconceived (if, perhaps, not yet thought out or articulated) attitudes and beliefs. While a juror's reasoning is ideally interchangeable with that of any other juror, the opinions of a focus group participant should be personal and idiosyncratic, generated according to a ‘natural,’ not an ‘official,’ logic of opinion-making. For it is only through the adoption of an everyday repertoire of judgment and talk that the desired liveliness of the focus group can be achieved. ¹³

The expected idiosyncrasy of individual opinions raises a fundamental practical problem for the moderator: the matter of influence. Not, as in the case of direction, the influence of the moderator herself on ‘the informant's own expression of feelings,’ but rather the influence of other focus group participants on the authenticity of individual opinions. After all, focus groups bring subjects into contact with one another on the assumption that the dynamic of their interaction will favour the generation of individual opinions. Yet, at what point does the group dynamic produce opinions that are no longer genuine and individual but simply the effect of processes of influence among participants? When is the influence of some members, or of the group as a whole, significant enough to render the judgments expressed by its members inauthentic? In other words, when is the product of a focus group simply a ‘focus group discussion,’ a series of opinions produced for and by the occasion, and not a series of genuine viewpoints? As one researcher puts it:

These are all scenarios in which the dynamic of the focus group gives rise to misleading objects: insincere beliefs, induced judgments, opinions produced under the influence of others or only for the purpose of the research at hand. Some critics see in this quandary an insoluble paradox, a contradiction in terms that fundamentally limits the usefulness of focus groups as a valid instrument of investigation: evidence that ‘what is often witnessed [in focus groups] are group attitudes and not the individual expression of an attitude’ (Morrison, 1998: 185)

To proponents of focus group research, however, the problem of influence is not an impossible contradiction at the heart of their methodology, but in fact simply an issue of ‘quality control,’ to be effectively addressed by the managerial skills of the moderator. ¹⁴ The techniques in which moderators must be proficient are precisely ‘those that seek to maximize the benefits and minimize the limitations of group dynamics by properly controlling them’ (Greenbaum, 1998: 110).

Yet, the term ‘quality control’, with its technical undertones, can be misleading, for what moderators articulate in their reflections on the problem of influence is not simply a purely technical understanding of ‘quality,’ but their implicit political philosophy. Discussions of the problem of influence are the touchstone on which moderators formulate an image of the social order most conducive to the expression of authentic individual opinions. By discussing the counterproductive aspects of the group situation, they provide a procedural definition of the ideal ‘small moral world’ of a focus group, a space of talk in which contradictory forces are kept in an artificial balance so as to generate the conditions for a fruitful exchange of genuine individual opinions.

Broadly speaking, and allowing for multiple differences of nuance and emphasis, the focus group is designed (and the moderator trained) to generate an isegoric situation. Classical Greek thought described isegoria as the condition of equality in the agora, understood as equality in the ability to express one's own opinions. Isegoria would not describe what we might understand today as ‘freedom of speech,’ the liberty to say whatever is on one's mind (that ‘freedom’ could rather be described, as I will argue below, as the virtue of parrhesia, which is often uncomfortably close to the vice of loquacity, an excess of speech that sometimes expresses a lack of authenticity). Isegoria refers to the formal conditions of an assembly in which citizens would have an equal share in the debate in the agora; it describes the quality of a space in which every member of the community is granted the right and the obligation of deliberative participation. ¹⁵

In the focus group, the research subject does not enjoy anything resembling ‘freedom of speech,’ for his or her speech is constrained by the objectives of the research and limited to the topics of discussion presented by the moderator. But the moderator nevertheless aspires to make the proceedings isegoric, in the sense that every research subject is not only allowed, but enticed and incited to form views and express opinions, and that any tendency towards a monopolization of the powers of argument (let alone overt intimidation) is strictly curtailed. ¹⁶ Using an altogether different metaphor, but expressing a similar view, the role of the moderator in bringing about isegoria is also described in terms of ‘diplomacy’:

Even excesses in rhetoric need to be limited in the focus group. While arguments often need to be drawn out to their ultimate consequences, rhetorical persuasion is dangerous because it tends to produce ‘group attitudes,’ and not ‘the individual expression of an attitude’ – a collective, rather than a collection of individual opinions. In the focus group, the moderator addresses the need to strip statements of their rhetorical quality through a series of directive techniques. As Puchta and Potter have shown in their analyses of focus group discussions, moderators visibly ignore – for instance by not writing down – statements directed by one participant to another, and persistently redirect participants to speak to the moderator, so as to generate ‘freestanding opinions,’ rather than contextually specific responses to other participants’ statements. Moderators ‘display attention (that is, visibly attend) to freestanding opinion formulations and display disattention to (explicitly) rhetorically embedded formulations’ (Puchta and Potter, 2002: 351). By inciting non-rhetorical formulations of participants’ opinions, moderators not only try to generate formally egalitarian conditions of deliberation. They also anticipate the need to produce a retrospective account of the focus group discussion stripped of its contextual specificity, an account that often relies heavily on the reproduction of illustrative quotes, which can be circulated far beyond their site of production and signification (Krueger, 1994: 167; Puchta and Potter, 2002: 360).

The ability of moderators to manage processes of influence, rhetoric and persuasion among participants should remind us that, in the focus group, isegoria among research subjects is compatible with an unequal distribution of the powers of speech between subject and moderator. In fact, from the point of view of the moderator the achievement of this equality in speech is simply the result of his own ‘isegoric skills,’ of his or her ability to maintain the evenly distributed liveliness of the interaction by preventing the concentration of talk and influence. We have seen in previous sections the importance that moderators grant to maintaining their authority vis-à-vis the group, their emphasis on the need to adopt a ‘style of leadership’ that would guide the subjects toward a productive exchange of opinions. The focus group is in this sense closer to the confessional or the pedagogical models of elicitation than to the ideal of the democratic assembly. The goal of the moderator is to benevolently (forcefully, yet imperceptibly) lead the focus group to a useful outcome (of which their subjects are ignorant). And, as I noted before, a fundamental objective of the moderator's techniques is to prevent the emergence of a ‘collective’ out of the group dynamic – to protect the individuality of the opinions expressed by each participant. ‘The moderator,’ one of them argues, stressing this point, ‘is a bit like a puppeteer, controlling the action, yet hoping panelists don't see her pulling the strings’ (Goebert, 2002: 35).

The opposite of an isegoric assembly is one in which some participants remain silent, and it is these silences – whatever their cause – that moderators aim to dissolve. ‘The facilitator,’ a moderation manual states, ‘must not just avoid domination of the group by individual members, but must also seek to encourage contributions from the most timorous’ (Bloor et al., 2001: 49). Thus, in the focus group setting isegoria implies that the elicitation of speech among participants must be evenly distributed, and that every member of the group must be given the encouragement and opportunity to express his or her opinions.

The logic behind this organization of talk is a purely utilitarian one. If elicitation is not evenly distributed, ‘valuable insights may be lost.’ In other words, any and every insight is potentially valuable to the moderator and her clients. ‘The focus group,’ one moderator writes, ‘rests on the deceptively simple premise that consumers can impart valuable information’ (Goebert, 2002: 32). Not only that: every participant is a potential source of value. If the moderator is to fulfill his productive mission, he ‘must truly believe that the participants have wisdom no matter what their level of education, experience, or background.’ (Krueger, 1994: 101) The ‘deceptively simple premise’ that value can be extracted from people by inciting them to formulate views and judgments, explains why the economy of discourse of the focus group is one of almost unrestricted proliferation. After the focus group meeting the researcher will have to code the multiplicity of opinions and screen out those considered irrelevant or inauthentic, distilling the most important trends into a report that quotes or paraphrases the opinions considered most representative or illustrative. But the larger the number of opinions expressed by the research subjects, the more likely that moderators and researchers will be in a position to mine valuable insights from the discussion. This is why alternative opinions must be encouraged and dissent actively promoted. The moderator continuously ‘heads off premature agreement’ between participants and works against their tendency to avoid direct disputes (Myers, 2004: 126). ¹⁷ The goal is to maximize the power of the focus group to generate multiple and heterogeneous viewpoints.

Given this economy of speech, it is not surprising that truthfulness, as a characteristic of the statements produced by participants, is strikingly absent from the moderators’ reflections. Judging from their technical literature, opinions cannot be true or false – they can either be genuine or induced, frank or contrived. In this sense, the focus group is not only isegoric, but also a space of parrhesia, a form of speech in which ‘there is always an exact coincidence between belief and truth’ (Foucault, 2001: 14). Parrhesia is often translated as ‘free speech,’ but etymologically it simply means ‘to say everything.’ In the focus group, the goal is to elicit a complete record of the participant's own ‘frame of reference,’ not to ascertain the ‘objective’ quality of his views. The function of the moderator is to turn every participant into a parrhesiastes, in Foucault's words:

Foucault points out in relation to parrhesia that ‘truth-having is guaranteed by the possession of certain moral qualities.’ In the focus group, true opinions – that is, opinions that can be circulated as authentic and actionable – are guaranteed by the moral qualities of the moderator, and by the proper ordering of the research subjects and their discussion. This is a curious reformulation of the Socratic method, in which a good moderator, armed with the proper techniques of interrogation, is able to extract truthful opinions from the assembled research subjects – where ‘truthful’ means, once again, a correspondence with the subject's authentically held beliefs.

I have so far argued that in conducting a focus group, moderators must grapple with fundamental questions of political philosophy: the right style of ‘leadership,’ the adequate form of authority (and the limits to its exercise), the role and dangers of rhetoric, the mechanisms of elicitation, silencing, and exclusion. All these issues come to the fore in the moderators’ discussions of the problem of influence. We can analyze the focus group as a sort of laboratory polity – an experimental and transient community in which a particular notion of the proper social order must be instantiated. The focus group is in this sense a practical application of political philosophy, and the moderators’ reflections on good practice represent the articulation of an ideal social order. This social order is isegoric, rather than egalitarian: it grants every member of the assembly an equal opportunity to express his or her views but their talk is always steered towards the goals of the moderator. The ‘deceptively simple premise’ underlying this practice is fundamentally utilitarian: the assumption that if a larger number of opinions can be extracted from research subjects, the moderator will gain more valuable insights into the modes of judgment and behaviour of consumers. Over and over again moderators emphasize this fact: that the purpose of the focus group is to extract value from the discussion with research subjects. In the microcosms of the focus group, then, political organization and knowledge production are one and the same thing. The focus group is an instrument of knowledge production, but also a ‘small moral world’ that must be properly ordered so as to maximize this function.

One-way mirrors and the focus group chain

The philosophies of moderation articulated by moderators undergo their materialization in the choice and arrangement of the particular physical setting in which the focus group takes place. As noted above, focus groups are examples of ‘white room’ settings – well-demarcated, closed spaces, designed to facilitate the interaction of research subjects and their observation by researchers, while isolating all of them from external influences. The design of this locale, and of the artefacts with which it is furnished, is part and parcel of the focus group methodology. In fact, the lack of adequate facilities is a major challenge to the usefulness of focus groups as a reliable instrument of knowledge production, and moderators sometimes complain that ‘the weakest link in the focus group chain is probably the facility.’

In the last two decades, however, as the use of focus groups has become widespread in commercial market research, the ‘physical plant’ of focus groups has improved a great deal. ‘What was a mom-and-pop operation has become big business,’ an American researcher notes. ‘A number of corporate chains are opening up more and more facilities around the country. Modern facilities are often large, well-appointed, and even glamorous’ (Langer, 2001: 54–55). The ‘upgrade’ of facilities has not been an even process, and the peculiarity of the setting often reflects divergent methodological choices and a differential access to resources between professional moderators, who use the focus groups for market research, and those, mostly social scientists, who employ it for scholarly purposes. Users of the focus group for marketing research often criticize the material arrangements of social scientists, and vice versa. Sometimes, this is combined with different ‘national styles’ in the conduct of focus groups. An American market researcher describes as follows the kind of facility often used in other countries:

The focus group setting encompasses thus a variety of possible sites, and the variability is a reflection of vernacular methodologies, the level of material resources, and even national preferences.

Perhaps the most identifying and controversial element of the focus group setting is the mirrored window that separates the research subjects and moderators from the back room so that clients and researchers can observe the proceedings while remaining themselves invisible. ¹⁸ The physical separation of those directly participating in the focus group discussion (moderator and research subjects) from those observing and recording it raises the question of the real-time and in situ communication between the moderator and her clients (or fellow researchers) in the back room. But it also generates broader questions about the relationship between moderators and clients, and the embeddedness of the ‘white room’ discussion in the ‘focus group chain.’

The real-time communication between clients and moderators is a perennial problem and a pervasive issue in the moderators’ literature. The old practice of passing written notes from the back room, by which clients used to communicate their reaction to the ongoing discussion, is almost unanimously rejected by moderators today. Such an obvious external intervention into the discussion ‘disrupts the flow of conversation’ and undermines the moderator's position of authority vis-à-vis the research subjects:

The alternative of having the moderator regularly leave the room to ask for instructions is even worse. In general moderators are keen to limit contact with clients throughout the course of the focus group. It is easy to understand that the moderator's strenuous efforts to become the benevolent ‘authority figure’ in the room can be easily upset by ‘overzealous clients,’ and generally by any intervention that makes visible to the research subjects the larger setting of which the focus group discussion is just a part. When instructions are passed from the back room, it becomes clear to the participants that the moderator is himself the object of observation and moderation by people hiding behind the mirror. The experimental features of the setting – the first- and second-order processes of observation to which participants are being subjected – become then glaringly apparent, perturbing the natural course of an experimental conversation.

This leads some moderators, particularly those using focus groups for social scientific research, to reject the presence of clients and invisible observers altogether. For these moderators, the increasing technological sophistication of the research apparatus, particularly visible in the United States, is a hindrance, rather than an enabling element.

Thus, discussions over the mirrored window and other physical arrangement of the focus group address, implicitly or explicitly, the relationship between moderators – the technicians of elicitation that produce the raw material of the research – and the clients and users of the knowledge being generated. An important dimension of this relationship is the theatrical aspect of the focus group. The mirror gives rise to a multidirectional game of observation and attribution in which research subjects, moderators, and invisible observers are entangled. As an American moderator puts it:

The danger of trying too hard to entertain the back room is often mentioned in the moderators’ technical literature. ‘I have observed moderators who spend a disproportionate amount of time during the session trying to be funny or clever for the basic purpose of generating a reaction from observing clients,’ one moderator notes (Greenbaum, 1988: 52). Yet the theatrical gaze is multidirectional and can be highly stimulating for the moderator. While the one-way mirror allows an invisible audience to observe the performance of the moderator and his subjects, it also reflects this performance back to himself, and to the people assembled in the ‘white room.’ ‘Somebody once asked me what I like best about being a moderator,’ one moderator remarks. ‘It's the entertainment value. The one-way mirror is a little like a proscenium arch. Part of the reward is the exuberance I feel in front of the mirror when things are going well or the anxiety when they aren't going the way the client – and I – thought they might’ (Goebert, 2002: 34).

The one-way mirror invites assumptions on both sides of the wall. Sometimes it gives clients a sense of immediacy to the minds of the research subjects, or so moderators think. ‘The most beguiling aspect of focus groups,’ one moderator argues, ‘is that they can be observed in action by clients and creative people hidden behind a one-way mirror. Thus, the planners and executors of advertising can be made to feel that they are themselves privy to the innermost revelations of the consuming public. They know what consumers think of the product, the competition, and the advertising, having heard it at first hand’ (Bogart, 1995: 67). At other times, moderators interpret the invisibility of the observers in the back room as a sign of shyness, or even fear, on the part of their clients – a perception that emphasizes the authority of the moderator beyond the ‘white room’ itself:

Yet, regardless of the motives attributed to the unobserved observers, moderators perceive and resent their presence – invisible as it might be – as a form of control. ¹⁹ In a certain way, the mirror makes real and evident (if also physically invisible) the dictum that ‘the moderator is the instrument in a focus group interview’ (Morgan and Krueger, 1993: 6). This idea may in theory be acceptable to facilitators, but it must also be made imperceptible during their encounter with the research subjects. The mirrored window gives a material form to this ambiguity.

Conclusion: the political constitution of market opinions

Despite being a key resource in social-scientific and marketing research, opinions are a neglected object of investigation. Researchers spend a great deal of effort explaining why people have the opinions they have, or using those opinions to explain other social phenomena, yet they have devoted very little attention to how something comes to be counted as an opinion, to the conditions of possibility for something to become an ‘opinion.’ Opinions are too often treated as unproblematic objects, unmediated expressions of people's beliefs or values.

The relatively unproblematic status of opinions is surprising, given the fact that knowledge and knowledge claims are routinely subjected to intense analytical scrutiny. We have sophisticated accounts of the manufacture of technical and scientific knowledge, for instance, and of the instruments and technologies through which it is created, certified and circulated. Opinions, as the example of the focus group hopes to make clear, also have an instrumental history. They are generated in a highly mediated fashion and through complex technologies of investigation, yet they do not seem to merit the same analytical treatment and are often addressed as if they somehow emerged from individual preferences. The famous distinction between episteme and doxa is maintained, albeit in a curiously inverse fashion: we have come to understand valid knowledge as an entity in need of an explanation, but the existence of a field of opinions still appears to us as a natural phenomenon, a function of actors’ beliefs.

In the focus group chain, individual opinions are manufactured as such – as peculiar entities, different from knowledge claims. They are treated as materializations of personal viewpoints and certified as expressions of individual beliefs. Among many other things, focus groups produce experimental representations of consumer attitudes and opinions on market products. The ‘focus group chain’ is, in its most refined and standardized form, a machinery for the elicitation of individual opinions and for their integration into marketing strategies.

This chapter has focused on the techniques of moderation and the self-understandings that moderators bring to bear on the task of generating those opinions. I have tried to discuss these techniques and self-understandings, not only as methodological prescriptions but also as a peculiar epistemology of moderation and, moreover, as instantiations of a particular notion of political order. The technical literature of moderators is infused with moral and political issues. It discusses authority, rhetoric, exclusion, influence, equality. The focus group is thus a kind of laboratory polity, an experimentally assembled and transient community organized to extract knowledge about people – particularly about their behaviour in the marketplace – from the opinions they express. And it can only do so if it endows the group with a particular political constitution. According to the moderators, this constitution is based on a personal, even artistic, style of authority and self-discipline on their part, an isegoric distribution of speech among the subjects, and a high degree of autonomy vis-à-vis clients and audiences. From the point of view of moderators, tradable opinions about the market are best manufactured in these ‘small moral worlds,’ where they can handle their subjects with a combination of conversational virtuosity, a skilful application of technique, and an isegoric political philosophy.