1The idea of dedicating an issue of the Revue française de science politique to the use of collective interviews arose from the observation that this data collection technique, which was totally marginal a few years ago in the field of French political science, is today increasingly being used. In 2004, when the slim volume on collective interviews [1] was published in the “L’enquête et ses méthodes” book series, references to the use of this technique in political science were rare. In 2009, there were signs of a shake-up as a number of theses and research projects using this technique became available. For us, this shift justified the organisation of a study day under the auspices of the French Political Science Association (AFSP). [2] We aimed to take stock of which particular topics and types of problem had turned to collective interviews as a means of conducting not just pure research in the political social sciences, but also applied research, including in the market research sector. This issue is a continuation of this event but only includes those contributions focusing on academic uses of the technique. [3] Things seem to have developed further since 2009 as additional evidence of the spread of this technique to political science circles and beyond can be found. As a result of both their number and their diversity, these studies constitute a critical mass enabling an initial assessment to be made at a time when the need to stimulate the debate on the methods used in political science is regularly highlighted. Today, the inclusion of a collective interview component in a research project is an increasingly common practice, even if, more often than not, they act as an ideal complement within a “multi-method” mechanism. They are also sometimes found in studies of an ethnographic nature, whether their use is planned – for example when it is a question of studying the system of influence between relatives [4] – or on a more ad hoc basis, when the collective dimension of the interview becomes the unexpected result of the field dynamics. [5]

2The situation has therefore significantly changed. Recognised in social psychology, [6] collective interviews have for a long time been excluded from the panoply of canonical methods in French political science, or even in French sociology. This statement should be qualified, however, since a notable exception exists in Tourainian sociology under the label of “sociological intervention”, which has for a long time now been using – in a specific form – what we might term collective interviews. This notable exception fully justifies the dedicating of one of the articles in this issue to a reflection on the expectations, experience and evolution of “sociological intervention”, the now essentially analytical purpose of which moves it closer to certain uses of collective interviews which retain a principle of reciprocity and of controlled intervention on the part of the researchers (see the article by Olivier Cousin and Sandrine Rui in this issue). The rest of this issue is based on studies in political science that have recently relaunched the use of collective interviews, without claiming to fully represent the range of such methods. [7]

3This introduction aims to highlight two mains points of discussion which can be determined from a review of recent uses of collective interviews. We will first discuss the highly flexible nature of this technique, referring, in part, to its origins and the manner in which it has been adopted, after which we will look at the difference between its uses in French political science and those – considerably more numerous – uses within English language social sciences. Arguing in favour of methodological reflexivity, we will set about making another type of comparison: on the one hand, an internal comparison between the different ways of implementing collective interviews, and on the other, an external comparison between collective interviews and other available techniques. In summary, the main aim of this issue is to advance the methodological debate, by providing the opportunity to assess, on the available evidence, the comparative advantages of collective interviews, such as they are implemented in practical terms by some researchers.

Flexibility and diversity of use

4It is not a simple task to characterise the way in which collective interviews are used, principally because of the highly flexible nature of the tool. A reader not familiar with the method will probably be confused by the range of recruitment possibilities (whether to form groups with people who know one another, or with those who don’t), of moderation (very flexible, low level of intervention, or confrontation of participants with stimuli and ranking activities, role plays, etc.), to say nothing of the different choices in relation to analysis (whether or not to code data, to undertake comprehensive analysis or focus on sequences that certain experts term “sensitive moments”, [8] quantification or strictly interpretative analysis, etc.). If we can agree on the fact that the lowest common denominator of collective interviews is that they produce discursive data within a collective framework, we must acknowledge that the way in which researchers turn this to their advantage varies greatly depending on the case in question. Some studies put interaction at the heart of their analysis, whereas others are first and foremost interested in the product of this interaction, leaving to one side the mechanics of production. Moreover, certain research projects seek maximum “naturalness”, while others engage with the experimental nature of the study.

5This flexibility stems principally from the tool’s origins and its subsequent adoption. We will not go into the history here of the emergence of what are commonly known as “focus groups” in English language social sciences, as this is now reasonably well known and documented in France. [9] The two illustrious patrons of the focus group (Merton, and secondarily Lazarsfeld) initiated this technique within the context of an applied research project. The flexibility of collective interviews probably owes a lot to this, to the fact that their use was initially characterised by objectives focused on producing rapidly established and easily communicable results, and that a proportion of those who use collective interview techniques remain positioned in the interspace between the market research sector and the academic research sector. One of the clearest indications of this high level of flexibility can be found in the absence of accepted terminology: [10] how do you choose between the terms “focus groups” – sometimes translated into French by the expression “groupes centrés” (centred groups) or “groupes focalisés” (focused groups) – and “entretiens de groupe” or “discussions de groupe” (group interviews/discussions); or between “entretiens collectifs” (collective interviews) and “groupes de discussion” (discussion groups)? Behind this linguistic vagueness lies a genuine conceptual vagueness; but although the competition between terms can itself be seen as an indication of the tool’s lack of standardisation, we can nonetheless distinguish some lines of differentiation. In English-language research, the label “focus groups” has more or less won the terminological battle. For our part, we consider that the term “entretien collectif” (collective interview) constitutes the most inclusive and relevant generic term. It includes interviews conducted with groups of people who know one another, which may be termed “entretiens de groupe” (group interviews) or “groupes de discussion” (discussion groups), [11] but also covers interviews that bring together people who did not know one another beforehand; [12] it furthermore encompasses “focused” interviews, for example centred on particular scenarios, [13] and which can, therefore, be designated as more or less directive “focus groups”.

6Without predicting the dynamics which might characterise future use, it appears that French political science has, for the moment, principally resorted to collective interviews in order to analyse discussion and the forms of argument that it produces. In brief, the renewed interest in political discussion stems from two movements which are a priori unconnected. The first is a renewal of interest in the study of personal influence, following in the footsteps of Lazarsfeld, [14] while the second is part of the debate on public deliberation and, in particular, on deliberation by the so-called masses, or ordinary people, which raises questions such as: How are opinions formed collectively? By what mechanisms does discussion bring about reasoning, justification and argumentation with regard to a public problem? What happens to deliberation when it escapes the institutional mechanisms for containing it?

7In Anglophone social sciences, recent thinking concerning the study of political discussion draws on work using quantitative surveys [15] recording self-assessments of discussion practices, and on ethnographic research based on the observation of conversations in public places, such as in local convenience stores, which are conducive to the development of routine forms of sociability. [16] Few studies, by contrast, have been conducted which build upon the pioneering works [17] which mobilised collective interviews as a tool for studying study political discussion [18] per se. In France, there have not been the same large-scale quantitative surveys of discursive practices [19] and the systematic ethnographic study of “ordinary” conversations is still rare. Reflections on “ordinary” political discussions often rely, in fact, on collective interviews.

8The asymmetric comparison between the abundant Anglophone uses and the French uses which are, by contrast, still small in number, serves therefore to highlight a form of disparity. Part of the success of focus groups in Anglophone social science, in reality, is due to the fact that they were promoted by feminist researchers, and experts in the sociology of health, sexual behaviour, or care, because they constituted a tool particularly well adapted to the study of dominated, minority, marginalised or deviant groups. Thanks to this technique, the domination effect is significantly reduced by moving from an individualising face-to-face encounter between interviewer and interviewee to a mechanism enabling the expression of a “collective” that tends to be in an autonomous position in relation to the interviewer. In other words, the collective interview enables a form of transfer of power, with the group taking the reins from the interviewer who becomes a simple “facilitator” or “moderator” (an “animateur” in French).

9The first uses of group interviews in French political science cannot really be considered in this perspective, in so far as they do not really exploit – or at least not explicitly or systematically so – the technique’s potential to study supposedly the most disadvantaged groups with regards to politics. For example, the technique has not yet been specifically promoted to study the politicisation of women, the working-class, immigrants, the younger generation, etc. – with the exception perhaps of the use of “sociological intervention”. The collective research to which we have contributed and which serves as the basis of our article in this issue came about, however, through a desire to take into account categories of the population (in our case, people from the so-called popular classes and of precarious status, often characterised by migratory trajectories) which, through lack of social resources, find themselves on the face of it distanced from politics. Our analysis leads us to suggest that the collective interview, subject to certain safeguards, [20] is perhaps a mechanism that is as much, if not better adapted to this category of the population than other techniques, in so far as it allows individuals to take a particular stand, which is more difficult to do when responding to a questionnaire or participating in an individual interview. The fact remains that French researchers working on “ordinary” or “popular” politics, on the politicisation of dominated, marginalised or minority groups are often sceptical towards the contributions of collective interviews. Yet it seems to us that the criticisms lodged against the collective interview need to be discussed on the basis of empirical testing, by opening up the “research kitchens”.

A plea for methodological reflexivity

10Methodological reflexivity remains a difficult, sometimes perilous exercise and not always a profitable one. It is, in fact, always risky for a researcher to present publicly his or her questions, let alone his or her doubts, regarding the impact of the methodological approach that he or she took to the data produced. [21] This issue does not aim to settle all the debates that exist within the small community constituted by those who use and are familiar with collective interviews; however it aims to untangle the threads of what they have in common and what their differences are.

11From observation, those who make use of collective interviews do not all analyse the same thing: some are interested in the “reasoning” (Pierre Lefébure), and the “judgements” and “argumentative repertoires” upon which they are based (Jean-Michel Lecrique, Pierre Lascoumes and Philippe Bezes). Others talk of “schemas” (Jean-Baptiste Comby) to emphasise the social roots and fixity, even the transferability, of the ways of envisaging a problem. And then there are those – including ourselves – who focus on how people adopt a position; the collective interview being used then as a mechanism for assessing the intensity of opinions. Users all agree however on the fact that the benefit of collective interviews is to put the relationship with politics back in a semi-public context by re-embedding the individual in a collective context, thus turning the spotlight on social influences [22] which are the basis of the mechanisms by which citizens position themselves in relation to the political order. The collective interview thus appears as a tool adapted to studying the co-construction, revelation, clarification and also the justification of remarks made within a collective context. Indeed, individual interviews and questionnaires have sometimes been criticised because they produce an individualised, even psychological, vision of how people relate to politics. This being the case, one advantage of collective interviews in relation to other techniques is, firstly, that they make social pressures both visible and analysable. Unlike individual statements in a questionnaire or an individual interview, collective interviews grasp the practice of discussion through action, which is neither reconstructed nor reassessed. Unlike ethnographic observations, they enable the researcher to go beyond merely collecting clues thanks to the magnifying effect that they create, and which proves particularly useful in the face of the difficulties that every “ethnography of citizenship” [23] study encounters. In fact, if quantitative surveys are to be believed, citizens (Americans in this instance, as these surveys were conducted in the United States) do indeed discuss public problems; however these discussions take place, in general, within politically homogeneous networks of individuals who reinforce each other’s opinions. Yet networks do exist where individuals are confronted by people who do not think like them; but here it is a question of “weak ties” [24] which bind work colleagues or simple acquaintances (friends of friends). [25] It is not easy, however, to find and observe places and situations within which political disagreements are expressed, and often one is forced to resort to forms of experimentation both in terms of recruitment methods (bringing together people linked by weak ties or who do not know one another) and in terms of ways of moderating the interviews (introducing political stimuli, exploiting disagreement, etc.).

12Among the points most debated by those who use collective interviews is the issue of what the comparative advantages are of the collective interviews that bring together groups of people who know one another in relation to those collective interviews that bring together people who do not knowone another. [26] The issue of the (dis)inhibiting effect that knowing one another has on what can be said within the group, in terms both of explaining oneself, but also of expressing heterodox ideas or deviant or stigmatised personal experiences, notably in relation to sensitive or controversial subjects, is at the heart of this debate. It appears that the space to express things which are easy, important, risky, indeed impossible to say, varies in a complex fashion, [27] depending on a number of parameters such as how sociable people are, the degree of familiarity existing between interviewees, how closely matched their social status is and, of course, the theme of the conversation; [28] doubtless other contextual elements also play a part, such as the location of the conversation for example. The issue of how far discussion dynamics are realistic when reconstructed in this way is clearly also raised. Configurations based on acquaintanceship have the advantage of bringing into play existing links. The fact remains that knowing one another often leads interviewees to leave things unstated, which means such things thus escape the interviewer who lacks the context to interpret what is only implied. In short, although they aim for a form of naturalness, these mechanisms are not, however, consistent with a genuine ethnographic perspective which would entail long-term observation and, of itself, allow insights into meticulously collected discourse. Configurations which bring together people who do not know each other must pay particular attention to recruitment (putting people together from the same social background, between whom relationships will be possible and plausible). [29] But above all, they must focus their analysis on the issue of the impact of the mechanism they have created and look at the processes by which the people assembled attempt to construct a shared world, which is a prerequisite for discussion.

13Beyond this, these considerations converge on the usual question: does a “truth” exist which would define the interviewees and which one research mechanism would be better placed to grasp than any of the others? The rare researchers who use collective interviews and have tackled this question recommend moving beyond the idea that a “methodological truth” [30] “in itself” exists. In line with a pragmatic approach, it seems more worthwhile to focus on how people justify their remarks depending on the situation. Different types of discourse can thus be expressed in public and private arenas, without there necessarily being any contradiction between them, nor with the degree of intimacy being objectively attributed a superior explanatory power for such discrepancies. Of greater interest, perhaps, is the shift from one to the other, without a priori deeming one to be truer than the other.

14With regard to the evaluation of different research techniques, as yet we have very few systematic comparisons. The comparison between research by survey and collective interview is generally made on an aggregate level. For example, in this issue, the mechanism put in place by Jean-Michel Lecrique, Pierre Lascoumes and Philippe Bezes enables comparisons to be made between the opinions expressed on public probity recorded by survey and those developed in a collective discussion. Although, overall, the structure of the opinions remains the same, the discussion sets in motion a moderating effect linked to the expression of forms of ambivalence. For our contribution, we compared responses to a questionnaire with the contributions made to discussions, both on an aggregate and individual level. Our findings indicate that collective interviews enable us to bypass the imposition of meaning which results from the standardisation and framing of responses to questionnaires, and to observe the ambivalence which characterises individuals’ opinions or the stands they take when they are invited to express themselves in a semi-public context. To us therefore it seems useful to focus energy on what collective interviews enable us to study: the strength or fragility, the stability or variability, the ambivalence or clarity of what is said, the common elements or disagreements which vary depending on the situation and the research model. Thus, we hope that this issue will open up new directions to those interested by this technique, enabling them to determine which methodological combination and which form of collective interview will be the most appropriate for their research strategies.