Do survey respondents always say the same thing?

Florence Haegel; Guillaume Garcia; Jean-Yves Bart

Do survey respondents always say the same thing?
Matches and mismatches between questionnaire responses and contributions to collective interviews
Suivre cet auteur Florence Haegel, Suivre cet auteur Guillaume Garcia, Translated from French by Suivre cet auteur Jean-Yves Bart
In Revue française de science politique Volume 61, Issue 3, May 2011, pages 483 to 511

format_quote Cite share Share Switch to French

1Political sociologists often disagree about the “right” method to use in order to comprehend ordinary relationships to politics, i.e., to analyse the way citizens – who are neither political professionals, nor activists, nor even particularly well-informed about politics, or interested in it – perceive or respond to politics. In recent years, French political scientists have made use of a broader range of research methods, principally via the increasingly widespread use of the ethnographic approach, thought to be more naturalistic, [1] and (more recently, and to a lesser extent) the use of collective interviews to study how opinions are constructed in a collective and interactional framework. [2] Although they differ, both these approaches proceed from the same basis: that opinions are co-produced, rather than developed by isolated individuals. They emerge from interactions and exchanges occurring in the context of discussions with family members, co-workers, friends, those who share our interests, or in the broader public sphere. This process requires appropriate observation techniques. Within the study of everyday politicisation, there is thus a plurality of methodologies. But in practice this means that academics spend much of their time discrediting techniques that they do not themselves use. And if such controversies often appear sterile, this is because very few credible attempts at comparing methods have been made. Indeed, as each researcher is generally only familiar with and defends one of these methodological approaches, there are few who spend time on the comparative evaluation of available techniques, [3] especially as such investigations can result in an unsettling form of methodological relativism for the researcher who undertakes it.

2We have nevertheless decided to take this risk by comparing questionnaire responses to contributions made in collective interviews organised within the framework of a comparative research project on attitudes towards European integration, conducted in Paris, Brussels and Oxford. Admittedly, there are many examples of “multimethod” research models which include the use of collective interviews, [4] and some discussion of the effectiveness of different techniques, and the congruence of the results they produce can be found in the Englishspeaking literature. [5] However, these studies generally do not question the same respondents, and thus do not include any individual follow-up which could clarify how questionnaire responses translate (or do not translate) into contributions to collective interviews, by situating them in the context of their production. We are dealing with a virtual methodological blind spot: although there have been a few rare publications in English on the subject, generally in the field of sociology of health or care, [6] the literature in French has completely overlooked the issue.

3In this paper, we will address our exploration (as yet incomplete) of this comparison by specifically analysing matches and mismatches between questionnaire responses and opinions recorded during collective interview discussions. This analysis should be seen as one step within a collective endeavour in which all members of our research team [7] are involved. [8] We are building on initial results collectively published by the whole team, [9] and on studies by Virginie Van Ingelgom on the ambivalence of opinions on European matters, [10] and by Sophie Duchesne and Florence Haegel on the politicisation of discussions. [11]

4The question we raise here, of an essentially methodological nature, is also relevant to European studies. Indeed, this field relies heavily on opinion surveys – a fact largely explained by the history of Eurobarometer surveys. [12] Though a “qualitative turn” has recently occurred, [13] most studies relying on techniques other than opinion polls still have to contend with the stronger legitimacy of quantitative studies based on questionnaires. In our team research, as we emphasised the Euro-indifference rather than the Euro-scepticism of the participants in our collective interviews, [14] we were clearly at odds with the results of those quantitative studies, which insisted on the rise of Euro-scepticism since the early 1990s [15] or the emergence of disputes between the losers and winners within European integration. [16] Our results, however, converged with other qualitative studies conducted at the same time, using other techniques or focusing on different populations. [17] Accordingly, in order to convince those of our colleagues who are traditionally wary of results based on a small sample and derived from the interpretation of particularly complex qualitative material, two strategies were available to us. The first consisted in reviewing Eurobarometer data in order to highlight, through renewed scrutiny of the evolution of its findings over time, elements validating the results derived from analysis of our collective interviews. [18] The second was to compare the stances taken by our respondents in collective interview discussions with their responses to two questionnaires administered during the phase of selection of participants and constitution of groups. [19] These interviews focus on European integration, but they also provide extensive data concerning the participants’ views on the economy and the market, the state, and immigration. This material can thus be contrasted with the participants’ responses to the questionnaires, particularly on Europe, the welfare state, and immigration.

5In our view, the specific advantage of collective interviews as we have used them lies in their ability to record stances rather than simple opinions – obviously as a result of their interactive nature. In other words, collective interviews record the intensity [20] of opinions; they can be particularly well suited to the expression of stances from members of the least socially and politically endowed categories of the population. Whilst establishing that perfect matches are scarce, we have nonetheless observed a significant amount of micro-discrepancies. These cannot be attributed to specific individuals, and therefore to purely individual or psychological factors; their frequency and type vary according to the context, the group dynamics and the issues discussed – each theme being characterised by a specific combination of social norms and experiences which the collective interviews highlight.

6Our argument is organised, in traditional fashion, as follows. First we discuss our research design – in many ways an atypical one. Secondly, we present, first globally and then individually, the results of the comparison of questionnaire responses and contributions to collective interviews, by contrasting European issues with the two other themes under study, the welfare state and immigration. Lastly, we will discuss the main methodological conclusions to be drawn from this comparison.

An atypical research design

7Our study is based on a comparative research project conducted in 2005-2006 on the politicisation of discussions, and focusing on the theme of Europe. That research relied on collective interviews bringing together previously unacquainted participants in three cities – Paris, Brussels and Oxford – in 24 collective interviews (N = 133). In relation to the traditional methods for organising collective interviews, our research design may be described as atypical given its experimental dimension, [21] evident in both the constraints placed on recruitment and on the approach taken to running the group. [22] Here, we will focus only on those elements that seemed essential both to understand the influence of the context in which participants were questioned, and to understand the conditions in which we were able to analyse and interpret the material.

Recruitment constraints

8From the outset we paid specific attention to the recruitment of participants, in order to avoid the widespread bias of most studies, which very rarely include sufficient representation of the socially and politically least endowed categories of the population. [23] To maximise our chances of mobilising members of the lower strata of the social hierarchy, we decided not only to compensate participants financially, [24] but also to get in touch with them through channels likely to reach them. [25]

9Besides the groups of activists, which we will not discuss in this paper, three types [26] of groups were constituted, following a tripartite modelling of social stratification wherein we distinguished what we call here upper, intermediate and lower groups. [27] This division, but also the relative internal socioprofessional homogeneity of each of the groups, met practical and theoretical criteria in two ways and were aimed as much at toning down social inequalities in the relationship to public speaking within each collective interview [28] as at taking into account what is already known concerning the social determination of attitudes towards Europe. [29] At the same time, we were seeking to organise discussions between citizens who disagreed on Europe, building on former research about conflictualisation. [30] To ensure this political heterogeneity, a questionnaire was administered over the phone, including a range of questions on socioprofessional characteristics, the general political leanings of respondents [31] and their position on the construction of Europe. [32] Subsequently, once respondents were selected, additional information on their sociographic profile, their political attitudes and their perceptions of European construction was collected through the administration of a second questionnaire. The latter included classic Eurobarometer questions; other questions, particularly on the welfare state and immigration, were meant to get a clearer grasp of the participants’ positions on a number of issues that are known to be ideologically structuring. With this information, we were able to compare (at least for most participants) their questionnaire responses and their contributions (or silences) in the collective interviews.

A specific moderation style

10Bringing together previously unacquainted individuals and following a specific scenography, [33] these collective interviews had decidedly non-naturalistic objectives. [34] Their moderation followed a similar logic. Within the main stages outlined in the scenario, the moderator [35] had two main goals. The first, a traditional one in collective interviews, was to ensure that contributions did not overlap too much, in order to make the recording listenable and the material interpretable. The second objective, a more unusual one (given that collective interviews are routinely used to seek shared meaning), was to facilitate the explicit expression of disagreements. Participants were asked to indicate incomprehension and disagreement by raising a picture of a flash of lightning, [36] and to steer discussion mainly towards potential points of disagreement. The script was also conceived to facilitate the expression of disagreements. After two initial phases focusing on more consensual questions about identification with Europe and the perception of the complexity of the European system of influence and decision-making, [37] we moved on to a potentially more conflict-laden second part of the discussion, including a question about the “winners and losers” of European integration (those who benefit and do not benefit from it) [38] and another about the accession of Turkey, preceded by a public vote. [39]

11It is important to note that our use of the collective interview here meant that it more closely resembled a test or ordeal. [40] Admittedly, as Gamson points out, [41] no collective interview can claim to accurately reflect the conditions of everyday conversation. Yet, the methodology we have used deliberately underlines the experimental dimension of the process. We have thus arguably subjected our participants to a “test of politicisation”, [42] whose objective was to observe how they reacted to it. In order to avert the risk of producing artificial effects, it seemed all the more necessary to refrain from being very directive in moderating discussions, and to allow the expression of the broadest range of possible reactions – including avoidance and diversion, or even silence. [43]

12Indeed, by placing people in a situation of co-presence, we create a set-up that amounts to a test insofar as each participant, by speaking, runs the risk not only of being contradicted and thus disagreeing with or even coming into conflict with others, but also of feeling inadequate or being led to reveal ignorance or confusion, of betraying a secret or holding a socially shameful opinion in a semi-public context. This is particularly the case for such “sensitive” subjects as politics. [44] We may therefore posit that participants are as a rule only willing to run such risks on subjects they deem very important and significant to them. Accordingly, this type of collective interview can rightly be considered as a technique for recording the intensity of opinions. [45] In other words, a position taken in a collective interview is the equivalent of a “mobilised opinion” in Bourdieusian terms [46] or “holding a view” [47] in Billig’s markedly different line of argument. [48] However, it is important not to reduce participant stances to the “expected” ones, in the sense that some of them derive from the position of participants in a given field. [49] Far from being mechanically reproduced, they are informed by personal implication and vary according to fluid dynamics of interaction. They relate to social experiences, but also to a concrete situation (the interview) which may sometimes be strongly normatively or emotionally laden. [50]

Coding the data

13For the purposes of this analysis, we only consider data collected in Paris and Brussels, excluding Oxford groups for practical reasons. We also exclude collective interviews conducted with activists, [51] as we focus on social comparison (without addressing the effects of party or union activism). Regarding themes, we have compared what interviewees say on Europe (on the basis of two items: the vote on the Treaty establishing a Constitution for Europe (TCE), and their opinion on EU membership), [52] on the welfare state (WS) [53] and on immigration. [54] Due to the different formats of the questions, we have not been able to compare responses term by term: questions on the TCE vote, EU membership or attitudes towards immigration offer the possibility of an equivocal (or “safe”) response, whereas the question on the welfare state is an either/or choice. This difference must be taken into account, and questionnaire responses were thus coded as follows (Box 1).

Box 1

Regarding Europe, two questions served as a basis for coding: the question on the (potential) TCE vote; and the judgement of the EU as a good or bad thing (or neither) for one’s country. Several combinations of responses were possible. Cases where responses to both questions were positive or negative were coded as “positive” or “negative”; contradictory responses (a positive and a negative one), “neither” or “don’t know” responses or refusal to answer one or both questions were coded as “ambivalent”. Cases where a positive or negative response was combined with a refusal to answer or a “neutral” answer were coded as “ambivalent positive” or “ambivalent negative”.
Regarding the WS and immigration, only one question per theme was taken into account. The resulting coding was “positive”, “negative” or “ambivalent”. The latter category includes non-responses to the WS question, non-responses and the equivocal response (“many, but not too many”) to the immigration question.
With two questions on Europe taken into account and the possibility of a “safe” response on immigration, more responses were mechanically coded as “ambivalent” for both of these themes than for the WS. Overall, there were eighteen “ambivalent” responses on Europe, sixteen on immigration and six on the WS.

14Also worth mentioning is that while both types of questionnaires were generally administered over the phone before the group interviews were organised, a minority of additional questionnaires were administered face-to-face, just before the discussion started, and a handful after the discussion. [55] Given how few there were, we are unable to test the effect of the conditions of administration of the questionnaire on responses; accordingly, we have excluded those rare cases from our analysis.

15We also systematically coded the synthesis of the entirety of each individual’s contributions on the three themes under review (see Box 2) in order to be able to compare them with questionnaire responses.

Box 2

Contributions to the discussion were coded as positive, negative or ambivalent. [56] This coding was done iteratively. We first singled out individual contributions on each of the three themes, and then, looking at them as a whole, we globally assessed the individual’s position on each theme: positive, negative, ambivalent, ambivalent/positive, ambivalent/negative, few or no contributions. This process proved anything but easy. For instance, if an individual repeatedly took a clearly positive stance on a given issue but took a negative stance only once, we coded him or her as positive, unless the negative stance was very intense and expressed at a key moment in the discussion (such contradictory contributions justified a coding as “ambivalent”). Quite often, individuals did not take a clear stance on the issue at hand, or were not talkative. Their contributions assumed the form of uncommitted “opinions”, descriptive or third-person discourse. To achieve a relatively subjective global assessment, we thus needed to take into account all the individuals’ contributions and their general tone, as well as body language, facial expression, silence, and more generally, the context of the contributions and of the discussion. We did not undertake double coding to ensure the reliability of our respective interpretations. Coding criteria were however established jointly. For an overview of the issues pertaining to the systematic coding of contributions (particularly assessments) in collective interviews, see Virginie Van Ingelgom’s paper on the entire corpus collected during this research. [57]

16Obviously, equating questionnaire responses with contributions to a discussion raises a significant issue of comparability. On the one hand, we assess a general “attitude” measured on the basis of a single question, and on the other, we collect a sometimes very wide and diversified range of positions – or sometimes, conversely, we have very little evidence if the issue has not been addressed by the individual or the group as a whole. The questionnaire frames the issue in a specific and therefore reductive manner, while contributions to a discussion may frame each problem in multiple ways. For instance, regarding the welfare state, the questionnaire focuses on the opposition between the value of justice and the value of labour, even though we know that attitudes towards the welfare state cannot be boiled down to this opposition. [58] Conversely, interviews elicit a range of positions on the benefits of competition, risk-taking and individual accountability vs. regulation, stability and social protection. The same goes for the subject of immigration: in the questionnaire opinions are assessed on the basis of a single, fairly classic question about the “number of immigrants”, whereas collective interviews provide a much more varied range of information on the participants’ thoughts about multiculturalism, differentialism, discrimination, Islam, etc.

Do collective interview contributions match questionnaire responses?

17In this section, we will begin with a brief overview of the findings of our analysis of questionnaire responses, insofar as they confirm well-known patterns. We will then go on to establish types of matches and mismatches between questionnaire responses and contributions to collective interviews, first at aggregate level, then at the individual scale.

Confirmation of patterns of production of questionnaire (non) responses

Figure 1

Social comparison of TCE votes

Figure 2

Social comparison of opinions on EU membership

Figure 3

Social comparison of opinions on the welfare state

Figure 4

Social comparison of opinions on immigration

18 Overall, our analysis of questionnaire responses [59] confirms existing knowledge about the production of responses and non-responses (the latter should be considered as a category of response in its own right). [60] Unsurprisingly, Europe elicits much more significant social variation in non-response [61] than the welfare state [62] and immigration. Although the nonresponse rate is virtually identical regarding the latter two, it almost triples on the TCE vote, and doubles on EU membership from upper to lower groups. This is a first indication that participants in our study (especially the most underprivileged) are quite unfamiliar with European matters. Participants in lower groups differ from those in upper groups through their indifference or their difficulty in taking a position on the subject rather than their opposition to European integration. This finding confirms those made by Virginie Van Ingelgom, who specifically analysed so-called “non-polarised” individuals. [63] According to her analysis, out of the 133 individuals that make up the global sample of the participants in our study, 33 belong to this category, including 20 from the intermediary and lower groups. [64]

19Unsurprisingly, given the number of studies that have established a link between ethnocentrism and educational attainment, the upper groups, more socially and culturally endowed, are less intolerant towards immigrants than the lower groups. We have also observed that members of intermediary and lower groups are more reluctant to agree that the welfare state “makes society fairer”. Building on Feldman and Zaller’s [65] research on US citizens’ perceptions of the welfare state, we may posit that members of the most underprivileged groups have a harder time dealing with abstract categories and tend to have recourse to a form of pragmatism. In that sense, their ambivalence towards redistributive justice [66] might be all the stronger as this redistribution is thought by these groups to profit immigrants, fuelling “welfare chauvinism”. [67]

20As far as national comparison goes, fewer members of the Brussels groups refused to adopt a position than in Paris, regardless of the issue addressed (Figures 5 to 8). At this point in our investigation, it was not possible to decide which of three possible hypotheses accounted for this: a social one, related for instance to inequalities in educational attainment in the sample structure; a methodological one, related to variations in the conditions of the administration of the questionnaires (including a possible “researcher effect”); or a culturalist one, positing a difference in the ideological and cultural structuring of the two populations and accordingly of the communities they represented. [68] Predictably, given the results of frequent opinion polls, [69] Brussels groups were more supportive of European integration. Yet, the question of their hypothetical vote on the TCE reveals a latent opposition (Figure 5) expressed in the discussions, specifically in their denunciation of Europe as something imposed on them, something with which they have been inculcated, or even infected (see below). [70]

Figure 5

National comparison of TCE votes

Figure 6

National comparison of opinions on EU membership

Figure 7

National comparison of opinions on the welfare state

Figure 8

National comparison of opinions on immigration

Mismatches observed at aggregate level

21For practical reasons, [71] we reduced the sample respectively to an upper group, an intermediary group and a lower group per city, i.e. a total of six collective interviews for a sample of 33 individuals (16 in Brussels, 17 in Paris). [72] For each of these individuals, we systematically linked their questionnaire responses and their contributions to discussions on Europe, the welfare state and immigration, in order to spot mismatches between the two types of data (see Box 3).

Box 3. Coding matches between questionnaires and collective interviews

Six groups made up of 33 participants were analysed. Three themes – Europe, the welfare state (WS), and immigration – were taken into account. Observations pertain to the degree of discrepancy of the responses/contributions by individual (n = 33) and by theme (n = 3). Table 1 is therefore based on 99 observations. “Comparison” refers to the relationship between questionnaire responses and contributions to the discussions. Five categories were distinguished: correspondence, fluctuation, contradiction, concealment and revelation. Again, when in doubt, we have systematically chosen the category indicating a greater discrepancy, in order to be as thorough as possible in spotting all forms of mismatches. For instance, an individual coded as “positive” in the questionnaire but who contributed very little to the discussion, even if the tone of his contributions was slightly positive, was classified in the concealment category. An individual coded as “ambivalent” in both cases was classified in the correspondence category. In this respect, our approach clearly differs from that of Virginie Van Ingelgom, which focused on polarisation. [73]

22We were thus able to identify a certain number of scenarios: some attest to an evident contradiction (positions adopted tend to contradict the opinion recorded in the questionnaire); others are discrepant to a lesser extent and may be called fluctuations (positions adopted during interviews appear to be less clear-cut and more ambivalent than questionnaire responses); yet others are movements of revelation (the collective interview reveals positions that had not appeared in the questionnaire responses) or concealment (the interview provides no trace of the opinions more or less clearly recorded in the questionnaire).

Table 1

Social, national and thematic comparison of questionnaire responses and contributions to collective interviews

Correspondence Fluctuation Contradiction Fluctuation + Contradiction Concealment Revelation Total Social comparison Upper 15 5 5 10 8 3 36 Intermediary 13 11 0 11 1 2 27 Lower 18 3 1 4 6 8 36 National comparison Paris 24 8 4 12 6 9 51 Brussels 22 11 2 13 9 4 48 Thematic comparison Europe 17 8 3 11 3 2 33 Welfare state 12 7 2 9 11 1 33 Immigration 17 4 1 5 1 10 33 Total 46 19 6 25 15 13 99

Social, national and thematic comparison of questionnaire responses and contributions to collective interviews

23Previous studies comparing data collected in questionnaires and in collective interviews have emphasised a form of congruence between both survey techniques. For the most part, our own comparison confirms this reassuring finding, as we have not observed significant discrepancies between the two. Blatant contradictions are very few and far between (6 cases out of 99 observed). [74] Correspondences are by far the most numerous (in around half of the cases, 46 out of 99). However, between clear correspondence and obvious contradiction, there is a grey area of micro-mismatches that we will now examine.

24If we aggregate contradictions and variations, we observe more social [75] than national differences, as these types of mismatches are more frequently found in upper and intermediary groups than in lower groups. Admittedly, more members of lower groups refused to respond to the questionnaires, but this reluctance was not apparent during the discussion. Revelations, indeed, were particularly frequent from members of those groups. They attest to the fact that participants of lower groups found it easier to express a position during the interview than give their opinion in the questionnaire. This finding, which we will analyse further below, is methodologically particularly interesting.

25The theme of the welfare state, which elicited few non-responses in the questionnaire, generated fewer contributions in collective interviews (11 cases of concealment) and numerous mismatches (9 cases of fluctuation or contradiction). The fact that the theme was not explicitly addressed in the script [76] does not alone explain why it was overlooked, since the same process did not occur with immigration. This trend is particularly significant in upper groups, and secondarily in lower groups, while it is residual in intermediary groups. This provisional finding should be validated with a wider sample, in order to establish a potential specificity of intermediary social categories in relation to the welfare state issue, and to confirm whether they would more openly adopt positions than the other groups on questions they deemed particularly sensitive or controversial for them.

26Immigration is the theme that elicited the fewest contradictions and fluctuations, but unlike the welfare state, many revelations. The questionnaire offered the possibility of an equivocal response, since the interviewees could answer that there were “many, but not too many” immigrants. It is precisely the individuals who refused to answer or opted for that response, and were classified as “ambivalent” on the basis of the questionnaire, who took a position during the discussion: they generally revealed their opposition to immigration, mainly on economic grounds (competition from immigrant workers, the risk of businesses being relocated) or religious ones (incompatibility of Islam with Europe’s Christian heritage).

27Europe elicited neither concealment (unsurprisingly participants discussed it at least a little, given the script) nor revelations; in other words, the discussion did not reveal positions that the questionnaire might have played a part in concealing. However, like the welfare state but in different ways, Europe was characterised by types of mismatches between questionnaire responses and contributions to the discussion. Considering the high number of “ambivalent” responses to the questionnaire (see Box 1), this finding confirms that the discussion did not contribute to making positions clearer; conversely, it even occasionally muddied the waters.

Mismatches observed at individual level

28The high number of micro-mismatches observed is unsurprising insofar as collective interviews inevitably allow a broader and often nuanced range of contributions and explanations, which makes it easier to account for the complexity and the ambivalence of opinions; moreover, as we have seen, our own coding choices have tended to highlight them (see above). One question, however, remains. In Table 1, where all matches and mismatches on the three themes considered are aggregated, are there individuals who are either specifically consistent in their responses, or conversely systematically inconsistent on all themes? In other words, do certain individuals contribute to these forms of discrepancies more than others? The hypothesis of a purely psychological dimension of the phenomena observed initially seems plausible: some people could arguably be considered as structurally hesitant, ambivalent, or influenceable, regardless of the issue discussed, while others could be particularly steadfast in their convictions, or subversive or nonconformist, which would enable them to resist the influence of the group. Our analysis, however, does not confirm this psychological hypothesis. Out of the 33 individuals in our sample, only four are completely consistent (their contributions to the discussion match their questionnaire responses on all three themes); none of them contradict themselves or even fluctuate on all three themes. With more flexible criteria, almost half of the sample (15 individuals) is mostly consistent (at least two matches on three themes) and the remainder (18 individuals) fluctuate according to the issue under discussion.

29This finding confirms both that there are few unwavering individuals, and that the individual parameter alone does not explain this variability; in other words, it cannot be attributed to specific individuals but is found evenly over the entire sample. This also allows us to conclude that the issue discussed, combined indissociably with the group dynamic, is one of the key parameters.

30As we cannot focus on each individual case within the entire sample here, we have decided to focus on the extreme or deviant cases, hypothesising that these will enable us to shed light on the mechanisms at work [77] (Table 2). We will first examine cases of pure correspondence, i.e., individuals who are completely consistent in their answers regardless of the theme (four individuals), and then those who are in contradiction on at least one theme (five individuals, including one, François, who contradicts himself on two themes).

Table 2

“Consistent” and “contradictory” individuals

Individual City Group Thematic coding Consistent individuals Louis Paris Upper Europe: Amb – WS: + Immigration: Amb Aline Paris Intermediary Europe: Amb – WS: NR/NC Immigration: Amb Pierre Brussels Intermediary Europe: – WS: + Immigration: - André Brussels Lower Europe: Amb – WS: NR/NC Immigration: Amb Contradictory individuals François Paris Upper Europe: +/Amb – EP: -/Amb + Stanislas Paris Upper EP: –/Amb + Jean-Paul Paris Upper Europe: Amb +/Amb – Claire Brussels Upper Europe: +/Amb – Ming Brussels Lower Immigration: –/+

“Consistent” and “contradictory” individuals

In the “thematic coding” column, we have indicated the coding of questionnaire responses and contributions to group discussions for each of the individuals. For “consistent” individuals, the coding is by definition the same; for “contradictory” individuals, the first code pertains to the questionnaire response and the second to contributions to the discussion. “Amb” stands for ambivalent, NR for a non-response to the questionnaire and NC for the absence of contributions to the discussion. The symbols + and – refer to the positive or negative orientation of the responses and comments.

31Before we proceed, let us point out that being “consistent” according to our coding method does not automatically mean that the individual has very clear-cut opinions, since consistently “ambivalent” individuals are included in this category (such as Aline, André and, partly, Louis). Furthermore, these “consistently ambivalent” individuals are not all similar. In line with the hypothesis of a gendered relationship to public speaking, Aline speaks little, actually increasingly less often as the discussion progress (like the other woman in the group, she lets the men talk); on the rare occasions she does speak, her contributions change slightly over the course of the discussion (relatively supportive of cultural openness at the beginning, she is more critical of the economic effects of Europeanisation at the end of the discussion). Conversely, André [78] and Louis contribute often and fully express the complexity of their opinions – a complexity that could easily be related to the frictions generated by their social trajectories and profiles. [79] Only in Pierre’s case do we find a case where polarisation and consistency of opinions go hand in hand (see Box 4). He is structured ideologically by his political affinities with the CDH (Humanist Democratic Centre, a Christian democrat movement at the heart of the Francophone Belgian party system); by a Catholic socialisation which has steered him towards a form of humanism that is critical of economic competition and supportive of values of solidarity and protection; and by a fierce defence of local and national identifications that leads to the exclusion of any outsiders.

Box 4. The intermediary Brussels group

The group comprises five individuals: Jonathan (29, IT employee), Maria (40, employee in the logistics sector in the public broadcasting company RTBF), Tina (32, currently unemployed and training to become a schoolteacher), Michèle (26, employee in the food industry) and Pierre (54, foreman).
The conversation begins with a discussion of Europe, but soon moves on to the issue of pensions. In the first five minutes, the “pro-Europeans” are quick to speak out. Jonathan proposes a very descriptive definition of being European: “being on a territory regulated by Community agreements”; Maria mentions exchanges and openness, and Tina uses the family metaphor. Pierre then chimes in: “I think it’s a sentiment that exists, but they’re trying to impose it on us in – I wouldn’t say unacceptable doses, but anyway, it’s being imposed when it should be coming more from the people themselves. It exists because that’s the way it is, at the level of the individual we are caught in an offensive from the outside, that’s what I think”. The “pro-Europeans” put forward three types of arguments: geopolitics (to counter the US), openness (promoting exchanges) and fatalism (national identity was also imposed, but it’s always been like that). However, they quickly give up on defending European integration. Later, Pierre and Michèle face off when discussing the example of Pierre’s daughter, a physical therapist who, he has said, cannot find any work. Pierre claims that “you need help when you start” and that there isn’t work for everybody (he also strongly blames the relocation of businesses and competition from Eastern European workers). Michèle answers that “you have to take risks”. A rapid-fire exchange ensues; Pierre gets emotional. Later, both are again in conflict regarding the financial perks that come with being a European civil servant: Pierre is shocked and Michèle simply claims she would be happy in their shoes. This issue is addressed again later in the discussion: Maria says that “making money isn’t shameful”, “money has to move around” and denounces Pierre’s “Judeo-Christian” views on money. Michèle concurs, citing the example of her boss who earns much more money than her, but deserves it because he takes risks.

32Pierre, who is a significantly older man in a group where women are in the majority, is one of the rare individuals liable to contradict the hypothesis positing that, in order to take a stance in a collective interview, it is necessary to find allies; [80] indeed, he makes no bones about expressing an initial position that puts him in the minority, by bluntly denouncing the European “norm”. This European norm, seen as a prescription associated with predictive and evaluative expectations formulated by institutions (for instance, school) or actors identified as members of the elite, works as a recurring catalyst for the positions he adopts on the European issue. [81] However, while Pierre’s strong stance effectively stops other participants from expressing other pro-European positions, this mechanism is not repeated when it comes to the other issues. Even if his positions in support of the welfare state and against Eastern European immigration and the relocation of business are also quite explicit, he does not succeed, in those cases, in “silencing” contradicting positions, and conflicts arise – sometimes on a fairly personal level. This is further evidence that the individual variable does not subsume the entire dynamic of the discussion and that the European theme elicits less involvement than the other two.

33Let us now consider the other extreme cases codified as contradictions. Two main mechanisms – sometimes cumulative – can explain these situations where individuals take stances that contradict their questionnaire responses. The first mechanism relates to a contextual effect. A collective interview is a social setting in which moves towards sharing and collectivisation occur, generating – as in other situations – effects of “conformity”. [82] As a result, some individuals may be pushed to contradict their questionnaire responses under group pressure. This applies to the case of Ming, a young waitress of Chinese origin in the lower Belgian group – the only instance of a contradiction on the immigration theme. Having stated in the questionnaire that there are too many immigrants, she found herself in a room with participants who all (except for one) shared a history of immigration. Early in the interview, she asks Dona, who has just discussed her Polish and Italian origins, about her “integration” and takes stances in favour of cultural openness. Against Christophe, the only participant with no history of immigration and who repeatedly makes strongly xenophobic statements, she even ends up expressing solidarity with rioters in the French inner cities, and eventually votes in favour of Turkey’s accession to the EU. Subsequently, she reconsiders this choice under the pressure of the rest of the group, mostly hostile to the Turkish accession; [83] this is further evidence of how sensitive to the collective dynamic she is.

34A similar mechanism appears clearly to be at work with the three individuals coded as “contradictory” in the upper Parisian group (Box 5). From the beginning of the proceedings, this exclusively male group is characterised by a relatively jaded attitude towards the construction of Europe, and a mostly very critical perception of the national and European political classes – the latter is judged as unrepresentative, corrupt and subservient to the economic forces of the global market. Correlatively, this group expresses a critical view of business relocations and redundancies, of the CAC 40 and the financial markets, of stakeholders and generally of financial capitalism. As the discussion unfolds, participants depict a rift between the political and economic elite who benefit from the system on the one hand, and the alienated “people” on the other (a category of which they consider themselves members, even though they make it clear that they are far from the bottom of the social ladder), who suffer from the measures taken by the former. It appears that the turn taken by the discussion affects the stances likely to be adopted in relation to state intervention. Any conspicuous endorsement of economic liberalism at national level might sound like a justification for the negative effects of globalisation; accordingly, participants liable to defend liberal positions were led to censor such views. As such, these cases of contradiction also appear, in this group, as the product of an effect of conformity and a result of contamination between the debate on Europe and the debate on the welfare state.

Box 5. Upper Parisian group

In this group, François (42, IT and logistics manager) expresses a positive opinion of Europe in the questionnaire. Yet, very early in the discussion, he shifts towards a cynical, or even openly critical stance on the construction of Europe (he spontaneously mentions the Euro, a “huge scam” that “messed everything up”, referring to it ironically as “Euroleptic”). Only his sarcastic tone strikes a different note from the general atmosphere of the group. His hostility increases during the second phase of the discussion, when the role of the market in Europe is discussed. One of the participants mentions EU regulations and the role of experts, who impose absurd and useless technical measures on people instead of dealing with their “real” problems. François then relates one of the only references to personal experience of Europe made in the discussions out of the three French groups analysed. Speaking about his father, a retired small winegrower who now receives a grubbing-up premium within the framework of the Common Agricultural Policy, he describes this measure, which he claims did not exist before the EU, as “unthinkable”, or worse, lethal (“that’s committing hara-kiri, this was the guy’s passion”). He also deplores that Europe favours large operations over small-scale producers like his father. His assertive rejection of a certain Europe thus seems to be explicable in terms of the significant, almost vital injury that it has inflicted on members of his family. This rejection appears to come systematically with the mobilisation of a populist register, denouncing the wrongdoings of the “big guys” that hurt the “little guys”. He explains the result of the TCE referendum as a sociological cleavage: the “yes”, he claims, is the vote of the “well-off” whom he denounces, not the vote of the “poor”, a category in which he considers himself a member by default – even though he had stated in the questionnaire that he voted in favour of the TCE.

35François’s case also illustrates the existence of another key element accounting for the rare cases of contradiction: the tension generated by the mobilisation of references to negative social experiences. [84] The last case of contradiction we will now analyse also manifests itself through a tension between adhesion to a European norm and the mobilisation of references to largely ambivalent experiences of Europeanisation. Claire (51) had trained as a doctor but at the time of the interview was working as an administrator in a patient advocacy organisation. She is a practising Catholic and a CDH sympathiser. Her group was made up of six highly “pro-Europe” participants (only two participants were slightly more lukewarm on the subject). She said she would have voted yes to the TCE referendum and found Europe to be a good thing for Belgium. She mentions her professional experiences, two European experts she knows, her brothers who have a European network of friends and acquaintances thanks to the Erasmus programme, etc. Yet, during the discussion, she says “one doesn’t really feel European, you know”, to which Valérie immediately replies “maybe you don’t, but I do!”. Claire then retracts her statement, explaining that she used an indefinite pronoun (“on” in French) and that she was not actually speaking for herself. But she then goes on to express reservations about her “European experiences”. Regarding the field of health, with which she is familiar, she explains: “your average doctor does not have an idea of Europe”. During the course of the discussion, she mentions the problems faced by European doctors in establishing their practice, the monopolies in the pharmaceutical sector, the reduced reimbursement for dentists in order to comply with European standards.

36Without going into the debate on the effects of Europeanisation on European societies [85] and the practices of their social actors, we can simply note here that Europe distinguishes itself from the other themes not only through the scarcity of references made to experiences of Europeanisation, [86] but also through the fact that these references can generate tensions that in turn elicit contradictions in the stances taken during discussions.

Discussion of the findings

37The findings we have just presented primarily constitute part of the debate on the variability of opinions, but they also raise questions on the methodological benefits of collective interviews, and in particular on the way in which we have used them in this research project.

The debate on the variability of opinions

38Following in the footsteps of Lane [87] and Converse, [88] Zaller’s research [89] has largely explored the enigma of the instability of questionnaire responses. [90] Without getting too deep into the complex analytical model he proposed and the debates it inevitably provoked, [91] let us recall that it is based on three main parameters: the often-demonstrated impact of survey methods on the opinions collected; the ambivalent nature of many political opinions; and the importance of the degree of salience – both in terms of the individual and of the public debate – of the issue considered. Individuals, based on what he terms contradictory “considerations”, are fundamentally ambivalent – not because they lack considerations but because of an excess of considerations that contributes to disorienting them. [92] Zaller thus explains that when they have to produce an opinion in a questionnaire, respondents choose a response between these “considerations” based on the salience of the issue, in other words whatever happens to be “on top of their head”. Though he admits that not everyone is ambivalent to the same degree, and that ambivalence varies depending on issues, he strongly emphasises the generalisation of the phenomenon. [93]

39The fact that we recorded what respondents had to say via two different survey methods, as opposed to the same technique reproduced identically, means we are not strictly in line with Zaller’s work. Within a short time period (questionnaires were generally administered between a few days and a few weeks before the collective interviews were organised), and by comparing two survey methods including one that logically generates more ambivalence, we established that the variability of opinions is frequent but mostly on a small scale; in other words, strict matches and mismatches of opinions are exceptional. Like Zaller and others, [94] we consider that these discrepancies cannot be ordered according to their degree of validity (scientific order) or of honesty (moral order). Different contexts unsurprisingly generate different types of data, [95] to which multiple meanings can be applied. While these discrepancies observed by the researcher might not have the same meaning for the actor, who may very well not remember them or even be aware of them, they do not merely result from chance or contextual effects, and cannot be directly attributed to psychological variables. Indeed, behind those discrepancies, there are classic social mechanisms at work, also found in everyday life, such as effects of conformity generated by social pressure.

40The second point of discussion pertains to the differences observed according to the issue under consideration. Zaller, who specifically worked on the welfare state issue, openly admitted to relying on qualitative studies, including Jennifer Hochshild’s individual interviews on social justice. [96] Hochschild is not interested in the variability of opinions, but in the ambivalence that she claims results both from conflicts of values and from a tension between norms and prescriptions (what one must think) and personal or perceived experience (what one experiences or sees). When reading her interview citations, [97] it is clear that ambivalence does not merely result from the multiple “considerations” that are more or less superficially cobbled together when the respondents produce their answers, but rather from intense tension felt by the individuals not only at “the top of their head”, but more deeply, within their social trajectories and experiences. Hochschild’s conception of the forms ambivalence takes appears to us to be better suited to understanding what is at stake in all forms of discrepancies recorded in collective interviews. It entails a concomitant observation of norms, values – and the conflicts that sometimes oppose them – and social experiences, and especially of the tensions brought about by their occasionally problematic combination.

41Analysing the entire corpus of this study, Virginie Van Ingelgom [98] has highlighted a number of research perspectives that may explain non-polarisation on the European issue, including the existence of an ambivalent relationship both to the “European norm” and to experiences of Europeanisation. Our study appears to indicate that this varies depending on the theme discussed. For instance, we can posit that the welfare state differs from the European issue on the basis that references to social experience are more frequently mobilised, and because it doesn’t depend on the degree of distance from a diffuse social or even institutional norm (also likely differing from immigration issues in that respect). As Zaller shows, the discussion of the welfare state brings conflicts of values to the fore, but few are the individuals who actually make a clear choice between them. [99] They tend rather to juxtapose these values, address issues separately, and vary their statements according to the situation and the framing of the discussion, which explains the fundamental ambivalence of stances taken on that theme, including in the most underprivileged social categories. [100] Europe and immigration elicited fewer such conflicts of values, instead activating a distancing from a social norm perceived as overarching.

The methodological benefits of collective interviews

42Studies comparing data collected in collective interviews and individual interviews emphasise the influence of group norms on what can be expressed during the discussion. [101] These norms, constructed or actualised in the group discussion, are moreover influential to the extent that they contribute to framing statements subsequently made in individual interviews. [102] As we introduced disagreement-inducing elements into collective interviews, traditionally used to produce common norms, our use of collective interviews responds to a concern to go well beyond merely recording opinions. [103] Yet, our choice of encouraging disagreement, while it does bring about conflicts, does not turn the entire discussion into a fight. Quite the opposite: if conflict must be actively fostered, it is because the social order and conventions of discussion spontaneously tend to defuse it. [104] In effect, conflicts often failed to arise, particularly in the case of Europe, despite a range of incentives. In other words, collective interviews, even in the perspective adopted for the specific purposes of this research project, still function as an instrument for recording group norms and shared meaning. Furthermore, we may posit that this construction of a shared world is all the more visible as we bring together people who are not familiar with each other. [105] In such groups, the process of construction of shared norms and meanings probably becomes even more necessary and explicit for the participants.

43Additionally, the social context we created inevitably confronted participants with the elite – given the framing of the discussion and the profile of the moderators. This context has likely heightened a mechanism of demarcation. Arguably, the sharp criticisms of European integration along the lines of “we don’t benefit from it, unlike them” and “they impose it on us” were raised quite quickly and very conventionally – including, but not only, in the lower groups [106] – as a banner under which the people gathered in the room could rally, sometimes implicitly deterring those of the participants tempted to cross the line. In other words, if the European norm is so easily denounced as overarching or even alienating, it is because such distancing is part and parcel of the construction of the group’s common meaning. Collective interviews allow us to record collective norms, i.e. specific to the groups physically gathered; likewise, they allow us to gain a better understanding of the mechanisms through which these groups try to distinguish themselves from or even oppose what they perceive as the out-group, often embodied by elites. [107] In short, the questionnaire arguably records a positioning in relation to a diffuse social norm and crystallised in the “right answer” (being in favour of Europe, repressing one’s xenophobic feelings) encouraged by the individual interviewer/interviewee relationship. By contrast, collective interviews enable the expression of collective norms relating to concrete groups, which may break from or sometimes openly contradict those social norms.

44Our findings also indicate that collective interviews are less inhibitive than questionnaires for the least privileged categories: indeed, a number of individuals from these categories adopted positions in the discussion, although their questionnaire responses were more ambivalent. This finding tends to show that the collective interview technique – including our atypical and potentially demanding version of it – is better suited than questionnaires to the expression of the positions of the least socially, culturally and politically endowed. In this sense, our analysis confirms a recurring observation in the literature: face-to-face interviews are more inhibitive than collective interviews, where participants, even if they do not know each other, can create a sense of togetherness and resist others by mobilising the codes and norms of social groups. [108]

45* * *

46Ultimately, this exercise in methodological comparison has yielded mostly reassuring results, as it attests to a degree of congruence between research designs. However, it is also to some extent disconcerting, since it reveals the existence of significant cases of subtle discrepancies. Our attempt to provide elements of interpretation and discussion on the basis of partial results is obviously not comprehensive. To explore this further, a more systematic examination of our corpus is necessary: it entails not only broadening the sample, but also codifying contributions – in particular references to social norms and experiences. Closer attention should be given to the contextual parameters we have mentioned and to interactions, including the formation of alliances, which we know to play a key role in determining the stances taken. [109] Lastly, it would also be worthwhile to confront our findings on “political” opinions with those of the only literature available on the subject so far, which is focused on issues of health and care but also addresses other “sensitive”, intimate subjects.

47Regarding the specifically methodological issues related to the use of collective interviews, our investigation has outlined a number of research avenues. First, by establishing that the individual is not the decisive parameter that explains the variability of opinions, and that we should devote our attention primarily to the themes discussed and the contexts of the discussion, we move away from a psychology-oriented conception towards an analysis of the specific socio-political workings of the discussion (what is being discussed? with whom?). This initial finding suggests that we should conduct comparisons not only between national or social groups, but also between issues and public problems characterised by varied configurations of norms and experiences. Our research also provides evidence that tends to confirm that collective interviews are particularly useful to record the positions of the least socially and politically endowed categories. This suggests reconsidering those criticisms which dismiss this method as unsuited to the observation of ordinary or working-class politicisation due to its “artificial” character. [110]

Notes

[1]
Stéphane Beaud, “L’usage de l’entretien en sciences sociales. Plaidoyer pour l’entretien ethnographique”, Politix, 35, 1996, 226-57.Online
[2]
David Morgan, Focus Groups as Qualitative Research (Newbury Park: Sage, 1988), 12.
[3]
Notable among the rare attempts are Jean-Marc Weller, “Le mensonge d’Ernest Cigare. Problèmes épistémologiques et méthodologiques à propos de l’identité”, Sociologie du travail, 36(1), 1994, 25-42. However, he does not address the variation in opinions, but the discrepancy between discourse on practice and observation of practice.Online
[4]
See Jean-Michel Lecrique, Pierre Lascoumes and Philippe Bezes, “Classer et juger les transgressions politiques”, Revue française de science politique, 61(3), 2011, 447-82, which compares results of opinion polls and focus groups. For classic work in this area, see Evelyn Folch-Lyon, Luis De la Macorra, Bruce Schearer, “Focus groups and survey research on family planning in Mexico”, Studies in Family Planning, 12(12), 1981, 409-32; Michelle Saint-Germain, Tamsen Bassford, Gail Montano, “Surveys and focus groups in health research with older Hispanic women”, Qualitative Health Research, 3(3), 1993, 341-67.Online
[5]
For a presentation of studies combining focus groups with either individual interviews or qualitative surveys, see David Morgan, “Focus groups”, Annual Review of Sociology, 22, 1996, 129-52.Online
[6]
Victoria Ward, Jane Bertrand, Lisanne Brown, “The comparability of focus group and survey results: three case studies”, Evaluation Review, 15(2), 1991, 266-83; Gregory C. Smith, Susan E. Savage-Stevens, Ellen S. Fabian, “How care giving grandparents view support groups for grandchildren in their care”, Family Relations, 51(3), 2002, 274-81. The most recent reference echoes our concerns as political scientists more closely: Amber Wutich, Timothy Lant, Dave White, Kelly Larson, Meredith Gartin, “Comparing focus group and individual responses on sensitive topics: a study of water decision makers in a desert city”, Field Methods, 22(1), 2010, 88-110.Online
[7]
This team brought together researchers from Sciences Po Paris (Sophie Duchesne and ourselves), the Politics and International Relations department of the University of Oxford (Elizabeth Frazer) and the Centre for Political Science and Comparative Politics of the Université Catholique de Louvain (André-Paul Frognier and Virginie Van Ingelgom).
[8]
The authors would like to thank Sophie Duchesne and Virginie Van Ingelgom for their insightful comments, which enabled us to refine our argument. We also wish to thank Anne Cornilleau and the anonymous peer reviewers.
[9]
Sophie Duchesne, Florence Haegel, Elizabeth Frazer, Virginie Van Ingelgom, Guillaume Garcia, André-Paul Frognier, “Europe between integration and globalization. Social differences and national frame in the analysis of focus groups conducted in France, Francophone Belgium and the United Kingdom”, Politique européenne, 30, 2010, 67-106. Online
[10]
Virginie Van Ingelgom, “Intégrer l’indifférence: une approche comparative, qualitative et quantitative de la légitimité de l’intégration européenne”, Ph.D. dissertation co-supervised by Sophie Duchesne and André-Paul Frognier, Paris, IEP de Paris/Louvain, Université Catholique de Louvain, defended on 28 May 2010.
[11]
Sophie Duchesne, Florence Haegel, “La politisation des discussions, au croisement des logiques de spécialisation et de conflictualisation”, Revue française de science politique, 54(6), 2004, 877-909; “Avoiding or accepting conflict in public talk”, British Journal of Political Science, 37(1), 2007, 1-22, “What political discussion means and how the French and the (French speaking) Belgians deal with it”, in Ken’ichi Ikeda, Laura Morales, Michael Wolf (eds), The Role of Political Discussion in Modern Democracies in a Comparative Perspective (London: Routledge, 2009), 44-61.
[12]
Philippe Aldrin, “L’invention de l’opinion publique européenne. Genèse intellectuelle et politique de l’Eurobaromètre (1950-1973)”, Politix, 23(89), 2010, 79-101; “The Eurobarometer and the making of European opinion”, in Daniel Gaxie, Nicolas Hubé, Jay Rowell (eds), Perceptions of Europe: A Comparative Sociology of European Attitudes (Colchester: ECPR Press, 2011), 17-35. Online
[13]
See Sophie Duchesne (ed.), “L’identité européenne entre science politique et science fiction”, Politique européenne, 30, 2010; Céline Belot, “L’Europe en citoyenneté. Jeunes Français et Britanniques dans le processus de légitimation de l’Union européenne”, Ph.D. dissertation in political science, Grenoble, IEP de Grenoble, 2000; Juan Diez Medrano, Framing Europe. Attitudes to European Integration in Germany, Spain and the United Kingdom (Princeton: Princeton University Press, 2003); Ulrike H. Meinhof, “Europe viewed from below. Agents, victims and the threat of the other”, in Richard Herrmann, Thomas Risse, Marylinn Brewer (eds), Transnational Identities. Becoming European in the EU (Lanham: Rowman and Littlefield, 2004), 214-44; Michael Bruter, Citizens of Europe? The Emergence of a Mass European Identity (Basingstoke: Palgrave Macmillan, 2005); Adrian Favell, Eurostars and Eurocities. Free Movement and Mobility in an Integrative Europe (Oxford: Blackwell, 2008); Jonathan White, “Europe and the common”, Political Studies, 58, 2010, 104-22 and “Europe in the political imagination”, Journal of Common Market Studies, 49(4), 2010, 1015-36; Daniel Gaxie, Nicolas Hubé, Jay Rowell (eds), Perceptions of Europe.
[14]
S. Duchesne et al., “Europe between integration and globalization”.
[15]
Liesbet Hooghe, Gary Marks, “Post functionalist theory of European integration: from permissive consensus to constraining dissensus”, British Journal of Political Science, 39(1), 2009, 1-23.Online
[16]
Neil Fligstein, Euroclash. The EU, European Identity and the Future of Europe (Oxford: Oxford University Press, 2008); Hanspeter Kriesi, Edgar Grande, Romain Lachat, West European Politics in the Age of Globalization (Cambridge: Cambridge University Press, 2008).Online
[17]
Specifically, our results match with those of Jonathan White’s study on taxi drivers in the UK, Germany and the Czech Republic, also based on collective interviews: see White, “Europe in the political imagination”.
[18]
This strategy was adopted by Van Ingelgom in part of her Ph.D. dissertation, “Intégrer l’indifférence”, chapter 2.
[19]
This was also done by Van Ingelgom in her Ph.D. dissertation, “Intégrer l’indifférence”, albeit differently.
[20]
We could also use the term “salience” (Duchesne et al., “Europe between integration and globalization…”, 82, note 11), but we prefer “intensity” here to avoid the confusion with Zaller, who uses the concept of salience to refer to the fact that an issue is on the interviewee’s mind, that he or she thinks about it “spontaneously”.
[21]
For a French-language introduction to the various methodologies used in conducting collective interviews, see Sophie Duchesne, Florence Haegel, L’enquête et ses méthodes. L’entretien collectif (Paris: Nathan Université, 2004).
[22]
On recruitment, see Guillaume Garcia, Virginie Van Ingelgom, “Étudier les rapports des citoyens à l’Europe à partir d’entretiens collectifs: une illustration des problèmes de la comparaison internationale en méthodologie qualitative”, Revue internationale de politique comparée, 17(1), 2010, 131-63. For an initial presentation of the analytical method, see Duchesne et al., “Europe between integration and globalization…” and Van Ingelgom, “Intégrer l’indifférence…”, particularly chapters 3 to 5.
[23]
On opinion polls, see Patrick Lehingue, Subunda. Coups de sonde dans l’océan des sondages (Bellecombe-en-Bauges: Éditions du Croquant, 2007), chapter 6: “Les sondés ne veulent plus l’être, les échantillons représentatifs le sont de moins en moins”. On collective interviews with previously acquainted participants, see the self-selection biases pointed out by William Gamson, Talking Politics (New York: Cambridge University Press, 1992).
[24]
50€ (or £40 in Oxford) for three or four hours of participants’ time.
[25]
See Garcia and Van Ingelgom, “Étudier les rapports des citoyens à l’Europe à partir d’entretiens collectifs…”.
[26]
In each country, two groups were organised by category in order to neutralise potential specificities of the group dynamics. Below we provide information on the lower, intermediary and upper groups in each city (six groups overall).
[27]
The terms chosen to describe our group categories are purely descriptive and not related to a value judgement of any kind. Upper groups are made up of executives and members of the upper strata of intermediary professions, as defined in the socioprofessional categories of INSEE (French National Institute of Statistics and Economic Studies). Intermediary groups are made up of the lower strata of intermediary professions, office employees and the upper strata of employees in the commercial and service sectors. Lower groups are made up of workers and the lower strata of employees in the commercial and service sectors; they include numerous individuals in precarious employment or unemployed.
[28]
S. Duchesne, F. Haegel, L’enquête et ses méthodes…, 47.
[29]
Céline Belot, “Les logiques sociologiques de soutien au processus d’intégration européenne: éléments d’interprétation”, Revue internationale de politique comparée, 9(1), 2002, 11-29; Jack Citrin, John Sides, “More than nationals: how identity choices matter in the New Europe”, in R. Herrmann, T. Risse, M. Brewer (eds), Transnational Identities…, 161-85; Bruno Cautrès, Gérard Grunberg, “Position sociale, identité nationale et attitudes à l’égard de l’Europe. La construction européenne souffre-t-elle d’un biais élitiste?”, in Olivier Costa, Paul Magnette (eds), Une Europe des élites? Réflexions sur la fracture démocratique de l’Union européenne (Brussels: Éditions de l’Université de Bruxelles, 2007), 11-35.
[30]
Duchesne and Haegel, “La politisation des discussions, au croisement des logiques de spécialisation et de conflictualisation”; “Avoiding or accepting conflict in public talk”; “What political discussion means and how the French and the (French speaking) Belgians deal with it”.
[31]
Evaluated on the basis of self-assessment on an ideological scale comprising ten categories (inspired by Eurobarometer surveys), and on how participants had voted in the first round of the latest general elections.
[32]
We asked French respondents how they had voted in the referendum on the European constitutional treaty, and Belgian and British respondents how they would have voted if such a referendum had been organised in their country. As the meaning of the question was clearly not the same for different national types, the comparison of the responses must be carefully monitored here.
[33]
Participants met in a location removed from their everyday surroundings (a place potentially labelled as “political”, such as Sciences Po) and talked in rather unfamiliar conditions (in the presence of a moderator, forming a semicircle), whilst being – visibly – filmed and observed by third parties.
[34]
The other difference with many ethnographic practices is that, as discussions are recorded and transcribed, the data collected and analysed is virtually exhaustive and therefore more precise.
[35]
In Paris and Brussels, groups were moderated by Sophie Duchesne and Florence Haegel.
[36]
This flash was then displayed on a board facing the participants, on which the moderator stuck cards outlining the key points of their contributions.
[37]
The suggested basis for discussion was: why is it or is it not a good idea to give power to nations, to elected representatives, to experts and to the market?
[38]
Participants responded first by sub-groups of two or three individuals, then in plenary sessions.
[39]
The discussion ended with an overview of the main national political parties’ positions on this issue. Each phase lasted between 30 and 45 minutes.
[40]
Even though our work does not share their approach, it echoes some aspects of pragmatic sociology: Luc Boltanski, Laurent Thévenot, De la justification. Les économies de la grandeur (Paris: Gallimard, 1991).
[41]
Gamson rightly points out that every collective interview, including those taking place in familiar surroundings with familiar people, yields a “social public discourse”: W. Gamson, Talking Politics, 19.
[42]
See Sophie Duchesne, Florence Haegel, “La politisation des discussions à l’épreuve de la comparaison: premiers enseignements d’une enquête en France, en Belgique francophone et en Angleterre sur le thème de l’Europe”, paper at the workshop “Regards croisés sur la politisation des individus: ici et là-bas, hier et aujourd’hui” organised par Mounia Bennani-Chraïbi et Jean-Gabriel Contamin, national conference of the Association française de science politique, Lyon, September 2007, http://www.afsp.msh-paris.fr/; S. Duchesne, F. Haegel, “What political discussion means and how the French and the (French speaking) Belgians deal with
[43]
François de Singly, “La gestion sociale des silences”, Consommation. Revue de socio-économie, 4, 1982, 37-63, esp. 48.
[44]
For a categorisation of so-called “sensitive” subjects including political themes, see Raymond Lee, Claire Renzetti, “The problems of researching sensitive topics”, American Behavioral Scientist, 33(5), 1990, 510-28.Online
[45]
Given, evidently, that not all contributions are stances or mobilised opinions. Part of the discourse recorded in our collective interviews amounts to a more distanced exchange of opinions; see S. Duchesne, F. Haegel, “What political discussion means and how the French and the (French speaking) Belgians deal with it”. Yet, the possibility of recording the intensity of opinions remains one of the additional benefits of this moderation technique.
[46]
Pierre Bourdieu, “Public opinion does not exist”, in Sociology in Question (London: Sage, 1993), 149-57.
[47]
Michael Billig, “The argumentative nature of holding strong views”, in Ideology and Opinions. Studies in Rhetorical Psychology (London: Sage, 1991), 168-94, esp. 171.
[48]
While Bourdieu connects stances taken to “the position one holds in a given field”, Billig relates them to everyday positions and interactions – for instance, in the case of collective interviews conducted with members of the same family, to the position and reputation one has within that circle. It’s worth noting that Billig works with collective interviews using a social psychology approach, which is not the case for Bourdieu.
[49]
P. Bourdieu, “Public opinion does not exist”.
[50]
See S. Duchesne, F. Haegel, “La politisation des discussions, au croisement des logiques de spécialisation et de conflictualisation”.
[51]
In non-activist groups, it was very occasionally the case that a participant referred to his or her involvement in a union or party (generally not a very active one). This was for instance the case of Farouk, a union representative in his company, which we will discuss later in relation to the lower Brussels group.
[52]
“Overall, would you say that EU membership is a good thing, a bad thing, or neither good nor bad?”
[53]
“Which of these opinions do you most agree with: ‘The welfare state makes society fairer’ or ‘The welfare state discourages people from working’?”
[54]
“Overall, what do you think about non-EU citizens who live (in France/Belgium/the UK): are there too many? Many, but not too many? Not many?”
[55]
Overall, four Belgian participants and nine French participants responded to the questionnaire face-to-face before the discussion began; six other French participants did so after the discussion. Out of the six groups which we will later analyse more closely, only one individual (a French woman) responded after the discussion.
[56]
It would be impossible to present here the list of arguments on the basis of which we assessed evaluations made by participants. A table synthesising all arguments mentioned by all participants in the study to justify their support or rejection of European integration is available in Van Ingelgom, “Intégrer l’indifférence…”, 198ff.
[57]
Van Ingelgom, “Intégrer l’indifférence…”, 193ff.
[58]
Stanley Feldman, John Zaller, “The political culture of ambivalence: ideological responses to the welfare state”, American Journal of Political Science, 36(1), 1992, 268-307, esp. 275.Online
[59]
Our sample breaks down as follows: (in terms of socioprofessional distribution) upper groups – N = 23, intermediary groups – N = 20, lower groups – N = 23; (in terms of national distribution) Paris – N = 36, Brussels – N = 30.
[60]
A non-response cannot “be considered as independent either from the respondent’s relationship to the subject of the question or from the actualisation of that relationship during the interview” (Fernando Porto Vazquez, “Les enjeux de la précision et du silence. Analyse d’un exemple des non-réponses à des questions de fait”, Consommation. Revue de socio-économie, 4, 1982, 13-35).
[61]
We have coded the following as non-responses: refusal to answer, no response, “don’t know”, blank votes, and “neither/nor” responses.
[62]
See Daniel Gaxie, “Des points de vue sociaux. La distribution des opinions sur les questions “sociales””, dans Daniel Gaxie, Annie Collovald, Brigitte Gaïti, Patrick Lehingue, Yves Poirmeur (eds), Le “social” transfiguré. Sur la représentation politique des préoccupations “sociales” (Paris: PUF, 1990), 141-92.
[63]
In other words, those who responded to the question on EU membership as a good or bad thing by “neither good nor bad”.
[64]
Van Ingelgom, “Intégrer l’indifférence…”, chapter 5 (251).
[65]
S. Feldman, J. Zaller, “The political culture of ambivalence…”.
[66]
Jennifer L. Hochschild, What’s Fair? American Beliefs about Distributive Justice (Harvard: Harvard University Press, 1981).
[67]
For a French-language presentation of the debate, see Virginie Guiraudon, “Des effets de l’ethnocentrisme sur les politiques redistributives”, Économie publique, 16(1), 2005, 3-12; for a recent analysis of the Dutch case, see Jeroen van der Waal, Peter Achterberg, Dick Houtman, Willem de Koster, Katerina Manevska, “Some are more equal than others. Economic egalitarianism and welfare chauvinism in the Netherlands”, Journal of European Social Policy, 20, 2010, 350-63.Online
[68]
Participants in the Brussels groups were members of the Francophone community.
[69]
According to the Eurobarometers, though their support is weakening Belgians remain more supportive of European integration than French citizens. See Van Ingelgom, “Intégrer l’indifférence…”, 98.
[70]
See S. Duchesne et al., “Europe between integration and globalization…”.
[71]
Even with such a reduced corpus, we had to list and categorise all contributions from each participant on the three themes considered, within data amounting to nearly 700 pages of text.
[72]
This choice could be debated on the basis of the comparability of the two “families” constituted for each social category (see S. Duchesne et al., “Europe between integration and globalization…”, note 4). We justified it by the fact that we wanted to select individuals with a low rate of non-responses on the questions we were addressing.
[73]
Van Ingelgom, “Intégrer l’indifférence…”.
[74]
This amounts to 6% of observations, which is close to what Ward and his colleagues observed (12%).
[75]
We noted a slightly higher proportion of women who did not take public stances or adopt more nuanced positions in interviews. In any event, the hypothesis of gender inequality in the relationship to public speaking should be validated over the entire sample and related to the group’s gender mix, a key variable when analysing women’s contributions.
[76]
Admittedly, since the script included a discussion of the accession of Turkey to the EU, immigration and the place of Islam, or even religion as a whole, were frequently discussed. But there were also opportunities to discuss the welfare state, when the respective roles of the market and the state in the European political system were addressed, or when the question “Who benefits from Europe?” was asked.
[77]
On the discussion and application of “case thinking” (pensée par cas), see Marie-Claire Lavabre, Florence Haegel, Destins ordinaires. Identité singulière et mémoire partagée (Paris: Presses de Sciences Po, 2010).
[78]
The fact that André is of Greek origin and does not speak very fluent French sometimes hindered the clarification and comprehension of his comments.
[79]
André is a precariously employed carpenter but studied until the age of 18; his political culture is alternative and ecological. Louis comes from a very right-wing family of naval officers but sees himself as left-wing; having failed the Sciences Po entrance exam and been unable to make a career as a naval officer himself due to a health problem, he became a part-time teacher and freelance photographer in his free time. During the interview, he provided much information on his personal experience in a group whose participants said little about themselves. André and Louis both strongly value the cultural dimension of Europeanisation.
[80]
See S. Duchesne, F. Haegel, “La politisation des discussions, au croisement des logiques de spécialisation et de conflictualisation “ and “Avoiding or accepting conflict in public talk”. It is actually not certain that Pierre’s case contradicts this hypothesis, as the authors of these papers mention that in some cases, allies are found outside of the group physically present – for instance in the cases of individuals active in associations, parties or unions. Even though we have no information on Pierre’s affiliations, the comments he made during the interview strongly suggest he might be involved in a local and/or religious association. In the other groups not analysed in this paper, other individual cases appear to be similar.
[81]
S. Duchesne et al., “Europe between integration and globalization…”; V. Van Ingelgom, “Intégrer l’indifférence…”.
[82]
Jocelyn Hollander, “The social context of focus groups”, Journal of Contemporary Ethnography, 33(5), 2004, 602-37.Online
[83]
She initially justifies her vote saying: “I put ‘yes’ because I suppose I don’t know the situation well, so I have a Turkish friend of mine who told me about it”; André (who, as we have mentioned, has Greek origins) then interrupts: “You should have thought twice!”, thereby exerting direct pressure on her.
[84]
The probability of the existence of such a negative personal experience must be taken in conjunction with the “atypical” character of the social origins of this senior executive, the son of a small winegrower. It is also possible that François’s responses to the questionnaire (he claimed to be against the welfare state) are informed by a stringent right-wing moralism rather than a genuine economic liberalism. In which case, the specific framing induced by the wording of the question would only allow us to record a limited fraction of the positions liable to be defended by him during the course of the discussion.
[85]
See Adrian Favell, Virginie Guiraudon, “The sociology of the European Union. An agenda”, European Union Politics, 10(4), 2008, 550-76.Online
[86]
See also D. Gaxie et al., Perceptions of Europe: A Sociology of European Attitudes.
[87]
Robert Lane, Political Ideology (New York: Free Press, 1962).
[88]
Philip Converse, “The nature of belief system in mass publics”, in David Apter (ed.), Ideology and Discontent (New York: Free Press, 1964), 206-61.
[89]
John Zaller, The Nature and Origin of Mass Opinion (Cambridge: Cambridge University Press, 1992).Online
[90]
John Zaller, Stanley Feldman, “A simple theory of the survey response: answering questions versus revealing preferences”, American Journal of Political Sciences, 36(3), 1992, 579-616 (583).Online
[91]
See for instance Paul M. Sniderman, Philip Tetelock, Laure Elms, “Public opinion and democratic politics: the problem of non-attitudes and the social construction of political judgement”, in James H. Kuklinski (ed.), Citizens and Politics. Perspectives from Political Psychology (Cambridge: Cambridge University Press, 2001), 254-88.
[92]
James H. Kuklinski, Part 3, “Introduction”, in Citizens and Politics…, 244.
[93]
P. M. Sniderman et al., “Public opinion and democratic politics…”, 256.
[94]
See for instance Jenny Kitzinger, “Focus groups: method or madness?”, in Matthew Boulton (ed.), Challenge and Innovation. Methodological Advances in Social Research on HIV/AIDS (New York: Taylor & Francis, 1994), 159-75.
[95]
In the case of this research, for instance, we could take into account the different contexts in which questionnaires were administered. While in most cases, the selection questionnaire and the additional questionnaire were administered over the phone, some were face-to-face, and a few were even self-administered for practical reasons. Moreover, the date of questionnaire administration was more or less close to the date of the interview. The length of administration, and the number of interactions necessary in the process, also varied. For all these reasons, the “considerations” that participants had in mind could change – this would require further investigation.
[96]
J. L. Hochschild, What’s Fair?…
[97]
See Jennifer L. Hochschild, “Disjunction and ambivalence in citizens’ political outlook”, in George E. Markus, Russel L. Hanson (eds), Reconsidering the Democratic Public (University Park: The Pennsylvania University Press, 1993), 187-210.
[98]
V. Van Ingelgom, “Intégrer l’indifférence…”, chapter 5.
[99]
S. Feldman, J. Zaller, “The political culture of ambivalence…”.
[100]
This ambivalence has been observed in the US and in France: see Jean-Luc Richard, “Les valeurs économiques: entre libéralisme et interventionnisme”, in Pierre Bréchon, Annie Laurent, Pascal Perrineau (eds), Les cultures politiques des Français (Paris: Presses de Sciences Po, 2000), 91-110. See also Richard Balme, “Convergences, fragmentation et majorités cycliques dans l’opinion publique”, in Pepper Culpepper, Peter Hall, Bruno Palier (eds), La France en mutation, 1980-2005 (Paris: Presses de Sciences Po, 2006), 375-422.
[101]
See Banks’ pioneering study where he interviews workers about organisational change in the workplace: John A. Banks, “The group discussion as an interview technique”, The Sociological Review, 5, 1957, 57-84. See also Daniel Wight, “Boys’ thoughts and talks about sex in a working class locality of Glasgow”, The Sociological Review, 42, 1994, 702-37; Lynn Michell, “Combining focus groups and interviews: telling how it is; telling how it feels”, in Rosaline Barbour, Jenny Kitzinger (eds), Developing Focus Group Research (London: Sage, 1999), 36-46.
[102]
Having conducted individual interviews before and after each focus group, Wight observed a greater gap between the two types of data when the individual interview was conducted before the group one; in the opposite case, the norm produced by the group continues to affect individuals once they find themselves alone with the interviewer. See D. Wight, “Boys’ thoughts and talks about sex in a working class locality of Glasgow”.
[103]
Experiments with polls have also been made to test this resistance of opinions: see Nonna Mayer, “La consistance des opinions”, in Gérard Grunberg, Nonna Mayer, Paul Sniderman (eds), La démocratie à l’épreuve (Paris: Presses de Sciences Po, 2002), 19-49.
[104]
See S. Duchesne, F. Haegel “Avoiding or accepting conflict in public talk”.
[105]
See J. Hollander, “The social context of focus groups”.
[106]
This is yet another illustration of the effectiveness of the “we vs. them” opposition: for a sociological perspective, see Richard Hoggart, The Uses of Literacy (Harmondsworth: Penguin Books, 1958); for a social psychology perspective, see Henri Taijfel, Differentiation Between Social Groups and Human Groups and Social Categories (London: Academic Press, 1978).
[107]
Leaving aside the out-group made up by “strangers” of all kinds.
[108]
The idea that collective interviews are well suited to collecting the opinions of dominated, minority or deviant populations is found across the literature, especially in a feminist perspective. See S. Duchesne, F. Haegel, L’enquête et ses méthodes. L’entretien collectif, 32; see also Jenny Kitzinger, “The methodology of focus groups: the importance of interaction between research participants”, Sociology of Health and Illness, 16, 1994, 103-21; Sue Wilkinson, “Focus groups in feminist research: power interaction and the co-construction of meaning”, Women’s Studies International Forum, 21, 1998, 111-25.
[109]
S. Duchesne, F. Haegel, “Avoiding or accepting conflict in public talk”.
[110]
This was one of the key elements behind the decision by the CURAPP research team in Amiens to set up a study day for researchers looking at ordinary relationships to politics in November 2010. We wish to thank the organisers, François Buton, Nicolas Mariot and Sabine Rozier, for giving us the opportunity to discuss our early findings on this theme.

English

Abstract

This article compares responses to questionnaires and contributions to focus groups carried out in 2005-2006 as part of research on attitudes toward European integration. It also compares the stances taken on Europe with those taken on two other issues, the welfare state and immigration. It therefore addresses questions about both the variability of opinions and the specific contribution of focus groups. It demonstrates a certain congruence between the two techniques as well as the existence of subtle shifts. These shifts are not attributable to individuals as such but to topics and to the contexts of discussion. This exploration provides evidence that focus groups are particularly useful to understand how citizens – and among them those with fewer social and political resource – are taking a stand.

An atypical research design
Do collective interview contributions match questionnaire responses?
Discussion of the findings
1. The debate on the variability of opinions
2. The methodological benefits of collective interviews

Florence Haegel

Florence Haegel is a director of research at the Centre d’études européennes (Sciences Po Paris). She is in charge of the “Political sociology and public policy” PhD programme at Sciences Po Paris. Recent publications include: (with Marie-Claire Lavabre) Destins ordinaires. Identité singulière et mémoire partagée (Paris: Presses de Sciences Po, 2010); (with Sophie Duchesne et al.) “Europe between integration and globalization. Social differences and national frame in the analysis of focus groups conducted in France, Francophone Belgium and the United Kingdom”, Politique européenne, 30, 2010, 67-106. With a team of researchers working on political parties, she recently co-edited the special issue “Enquêter dans les partis politiques. Perspectives comparées” of the Revue internationale de politique comparée, 17(4), 2011 (CEE, Sciences Po Paris, 28 rue des Saints-Pères, 75007 Paris florence.haegel@sciences-po.fr).

Guillaume Garcia

Guillaume Garcia is a doctor in political science and a research officer at Sciences Po Paris. His publications include (with Virginie Van Ingelgom) “Étudier les rapports des citoyens à l’Europe à partir d’entretiens collectifs: une illustration des problèmes de la comparaison internationale en méthodologie qualitative”, Revue internationale de politique comparée, 17(1), 2010, 131-63. He has worked on the media treatment of social movements and the European Union, and on attitudes towards European integration. He is currently in charge of coordinating the activities of the Banque d’enquêtes qualitatives at the Centre de données socio-politiques (Sciences Po Paris, 27 rue Saint-Guillaume, 75337 Paris cedex 7 guillaume.garcia@sciences-po.fr).

Translated from French by

Jean-Yves Bart

Latest publication on cairn or another partner portal

Gaël Villeneuve, Les débats télévisés en 36 questions réponses
Grenoble, Presses universitaires de Grenoble, coll. Politique en plus, 2013, 136 pages
In Questions de communication Volume 25, Issue 1, January 2014

Latest publication on cairn or another partner portal

Uploaded on Cairn-int.info on 03/03/2014

Cite article English French

ISO 690		Copy HAEGEL Florence, GARCIA Guillaume, "Do survey respondents always say the same thing? Matches and mismatches between questionnaire responses and contributions to collective interviews", Revue française de science politique, 2011/3 (Vol. 61), p. 483-511. URL: https://cairn:wghqyw8uhkb3@preprod.cairn-int.info/journal-revue-francaise-de-science-politique-2011-3-page-483.htm
ISO 690		Copy HAEGEL Florence, GARCIA Guillaume, « Les enquêtés disent-ils toujours la même chose ? Concordances et discordances entre les réponses à un questionnaire et les interventions dans un entretien collectif », Revue française de science politique, 2011/3 (Vol. 61), p. 483-511. DOI : 10.3917/rfsp.613.0483. URL : https://cairn:wghqyw8uhkb3@preprod.cairn.info/revue-francaise-de-science-politique-2011-3-page-483.htm
MLA		Copy Haegel, Florence, et Guillaume Garcia. "Do survey respondents always say the same thing? Matches and mismatches between questionnaire responses and contributions to collective interviews", Revue française de science politique, vol. 61, no. 3, 2011, pp. 483-511.
MLA		Copy Haegel, Florence, et Guillaume Garcia. « Les enquêtés disent-ils toujours la même chose ? Concordances et discordances entre les réponses à un questionnaire et les interventions dans un entretien collectif », Revue française de science politique, vol. 61, no. 3, 2011, pp. 483-511.
APA		Copy Haegel, F. & Garcia, G. (2011). Do survey respondents always say the same thing: Matches and mismatches between questionnaire responses and contributions to collective interviews. Revue française de science politique, 61, 483-511. https://cairn:wghqyw8uhkb3@preprod.cairn-int.info/journal--2011-3-page-483.htm.
APA		Copy Haegel, F. & Garcia, G. (2011). Les enquêtés disent-ils toujours la même chose : Concordances et discordances entre les réponses à un questionnaire et les interventions dans un entretien collectif. Revue française de science politique, 61, 483-511. https://doi.org/10.3917/rfsp.613.0483
DOI		Copy https://doi.org/10.3917/rfsp.613.0483

Download citation

file_downloadZotero (.ris) file_downloadEndNote (.enw) open_in_new RefWorks

Cite

Distribution électronique Cairn.info pour Presses de Sciences Po © Presses de Sciences Po. Tous droits réservés pour tous pays. Il est interdit, sauf accord préalable et écrit de l’éditeur, de reproduire (notamment par photocopie) partiellement ou totalement le présent article, de le stocker dans une banque de données ou de le communiquer au public sous quelque forme et de quelque manière que ce soit.

Full-text

An atypical research design

Recruitment constraints

A specific moderation style

Coding the data

Box 1

Box 2

Do collective interview contributions match questionnaire responses?

Confirmation of patterns of production of questionnaire (non) responses

Social comparison of TCE votes

Social comparison of TCE votes

Social comparison of opinions on EU membership

Social comparison of opinions on EU membership

Social comparison of opinions on the welfare state

Social comparison of opinions on the welfare state

Social comparison of opinions on immigration

Social comparison of opinions on immigration

National comparison of TCE votes

National comparison of TCE votes

National comparison of opinions on EU membership

National comparison of opinions on EU membership

National comparison of opinions on the welfare state

National comparison of opinions on the welfare state

National comparison of opinions on immigration

National comparison of opinions on immigration

Mismatches observed at aggregate level

Box 3. Coding matches between questionnaires and collective interviews

Social, national and thematic comparison of questionnaire responses and contributions to collective interviews

Social, national and thematic comparison of questionnaire responses and contributions to collective interviews

Mismatches observed at individual level

“Consistent” and “contradictory” individuals

“Consistent” and “contradictory” individuals

Box 4. The intermediary Brussels group

Box 5. Upper Parisian group

Discussion of the findings

The debate on the variability of opinions

The methodological benefits of collective interviews

Notes

Abstract

Abstract

Outline

Authors