How to approach ‘prejudice’ and ‘stereotypes’ qualitatively: The search for a meaningful way

Introduction

Over time, this article has evolved from a short, simple conference presentation focusing on a few interesting extracts concerning prejudice and stereotypes. When I began writing the article in “discussion” with other (cited) authors, looking for potential new ways of analysing the same extracts, it got far more complicated than I had anticipated. It now takes the form of part theoretical and analytical paper about stereotypes and prejudice, and part confessional tale (Van Maanen, 1988) of my journey. It is a journey that has been about looking for a meaningful and useful approach to analysing empirical material I have collected in different research projects over more than 15 years, in an attempt to say something about prejudice and stereotypes.

The study of prejudice became central to social psychology with the work of Allport (1954). Allport, who focused mainly on negative ethnic prejudice, defined it as “an antipathy based on faulty and inflexible generalization. It may be felt or expressed. It may be directed toward a group or an individual of that group” (Allport, 1954, p. 9). As Billig (2012) notes, prejudice was initially defined more broadly but at the beginning of the 20^th century it narrowed to refer to negative opinions and to focus on categories of ethnicity and race^[2]. Today (in mainstream social psychology) this concept can mainly be found in the triad of prejudice, stereotypes/stereotyping and discrimination; prejudice is usually defined more specifically as a complex attitude to a specific group, stereotypes/stereotyping as the attribution of specific characteristics to this group and discrimination as a non-neutral behaviour towards this group and its members (e.g. Dovidio, Hewstone, Glick, & Esses, 2010). Alternatively, this triad is understood slightly differently in the three-part model of attitudes (sometimes called the ABC model), in which stereotypes are seen as the cognitive aspect, prejudice as the affective aspect and discrimination as the behavioural aspect of attitudes towards a group (Fiske, 1998). Moreover, in recent writing numerous more specific concepts have appeared that have a more or less clear connection to the concept of prejudice and that can be used as dependent measures which tap into the ABC model. These include social distance, intergroup trust, perceived threat, and so forth (for an incomplete but rather voluminous and recent overview, see Lášticová & Findor, 2016).

Nonetheless there is immense variation in the social psychological work done on prejudice. As Condor and Figgou (2012) summarize more generally, in social psychology, prejudice has been studied as a matter of “instinct, drive, motivation, emotion, categorization, social identity, attribution, personality, executive control or rhetoric” (p. 202). There have also been many different opinions on the reasons for prejudice and the mechanisms by which it comes about. Dovidio (2001) describes how these have changed over time—the early works saw prejudice as a personality fault, later it was viewed as imperfect information processing and recently more and more researchers have focused on unconscious and automatic prejudices.

Again, when we look at how prejudice has been researched, there is great diversity. Very different methods were used at different times and by researchers working within different paradigms. Prejudice has been researched using qualitative (e.g. interviews, focus groups), quantitative (e.g. questionnaires, social distance scales), as well as experimental (e.g. pupil dilation, response latency) methods. The predominant methods, though, are perhaps those that are most easy to use: direct self-report questionnaires (for an excellent overview see Fiske & North, 2014). Usually those researched give their views (personal or for their whole group^[3]) on different groups defined for example according to race/ethnicity, age, gender, sexual orientation or other criteria, and researchers measure the attitudes of one group towards the other while working with different components according to the theory they subscribe to—competence, trust, warmth, and so on.

The role of the researchers is omnipotent here—they get to say what is (and is not) prejudice and whether and to what extent someone is prejudiced (or not); in short the researchers are the “standard setters” of truth (Kruglanski, 1989 in Dixon & Levine, 2012a, p. 306). As Durrheim and colleagues argue (2016), social psychologists have “generally sought to develop authoritative definitions and measures of prejudice” (p. 18) and these from top down definitions “have been superimposed on ordinary people’s attitudes in order to identify prejudiced individuals” (p. 18). Moreover, there is a problem with these definitions. As Condor and Figgou (2012) state, they are not very precise and have not exhibited a “high level of consistency over the past century” (p. 201). Despite social scientists’ claims that they have 1) been making the definition more and more accurate over time in line with scientific progress, and 2) that they are much more specific and precise than ordinary social actors—lay men and women—in fact their definitions are not so very different. However, as Durrheim, Quayle and Dixon (2016) state, in everyday communication what counts or does not count as prejudice is context dependent in the given situation and sometimes even fought over “with considerable passion and no little skill” (p.18), mainly because casting the same idea as prejudice or as a rational and legitimate attitude has very different consequences in the real world. Because of all the real-life complications with a clear-cut definition of prejudice, and because of other more generally critical voices (e.g. Whetherell, 2012), I have placed the terms “prejudice” and “stereotypes” in quotation marks in the title and also in important places in the article as a reminder that these are just labels.

Inspired by these insights and following my old suspicion of measurement tools, in this article, I am not interested in measuring the extent to which someone (a person or group) could be considered prejudiced, or how that changes after such and such an intervention^[4]. Neither am I interested in the content, what exactly the potentially prejudiced opinion is about or which characteristics are ascribed to the group of people in question. What I want to look at are the opinions of the participants, but I plan to approach prejudice from a meta position, so as to better understand what the participants say and possibly think ABOUT prejudice, rather than what their explicit definitions of it might be. Thus in this article I shall attempt an approach to stereotypes and prejudice where I am interested in how the research participants THEMSELVES refer to the existence and validity of prejudice/stereotypes (in themselves and others), what they think about how the stereotypes are shared within their own and other groups, how they personally (dis)agree with them, how carefully they express an opinion that could be socially unacceptable, how they work around this complication, and so on. All this is possible only in contexts where the participants can articulate their opinions (not for example in closed questions in a questionnaire). Thus the methods of empirical material collection already determine which approaches can and cannot be used with the material. Here I am using extracts from interviews, focus groups and (open limit) answers to an open question in a questionnaire. Another possibility would be to use statements that did not originate in a research setting, but that are “natural”^[5]—like newspapers articles or political speeches (see e.g. Wodak & Meyer, 2001).

I would like to make two small points before I present the empirical material collected and the analytical perspective: First, as mentioned above, implicit definitions of prejudice do not appear to differ so greatly in lay discourses and in scientific writing. However, in my experience, if ordinary social actors use (and often they do not) an explicit term when referring to something they consider prejudiced/one-sided/stereotypical, they tend not to distinguish between the concepts of stereotypes and prejudice but use both terms synonymously, or they sometimes use “prejudice” in reference to negative opinions only, and “stereotypes” for both positive and negative opinions. That is why I use both terms in this article, and, of course, I use the term the participants use in each extract from the empirical material. Second, in this article I deal only with nationally defined groups, often defined on the basis of state citizenship because this was the perspective from which our earlier research projects were conducted. This is despite my agreeing that nationalities are labels and that in real life issues around membership in nationally defined groups is often complicated, complex and not at all clear-cut (see also methodological nationalism, Wimmer & Glick Schiller, 2002).

Empirical material and analytical perspective

Over the last 15 years, I have been involved (in various teams with several colleagues) in a number of predominantly qualitatively oriented projects in which participants were divided up using national categories and which looked at what is happening in the borderlands or other areas where the members of these different national groups live their everyday lives together. We have never explicitly asked about “stereotypes” or “prejudice” as we were interested in other issues, but sometimes (not in fact very often) the participants referred to them explicitly or implicitly. Some time ago I decided to search all the different transcripts again in order to look specifically at those parts where this topic is thematized. The research projects are:

1. In 2003–2004 we used semi-structured interviews and commented drawings of the borderland in an Austrian–Slovak project about young adults from the borderland and their perceptions of their own nation and the other nation (32 Slovaks, 32 Austrians, aged 16–24, selected using quota sampling taking into account age, gender, education, size of dwelling and (not) having a better experience of the other nation) (see Spannring et al., 2005);

2. in 2005–2007 in a project on the lives, attitudes and feelings of home of Slovaks travelling regularly (mostly daily) to Austria for study or work we used focus group discussions, semi-structured interviews and commented drawings of the borderland (26 Slovaks, aged 24–46, selected using quota sampling taking into account age, gender, education and size of dwelling. We then looked for and added “contrasting cases”) (see Lášticová & Petrjánošová, 2014);

3. in 2009–2010 in a project on the everyday lives of a “community” of Slovak short-term migrants to Ireland we used semi-structured interviews with 8 (male and female) Slovaks who had different leadership roles in the Slovak group in Ireland (see Lášticová & Petrjánošová, 2013);

4. in 2010–2012 in a research project on intergroup attitudes in central Europe we analysed answers to an open question about experiences of the neighbouring nation. The respondents were Czechs and all their neighbours—Austrians, Germans, Poles and Slovaks (1,260 female and male university students from the respective borderlands) (see Graf, Hřebíčková, Petrjánošová, & Leix, 2015).

The theoretical perspective I adopt in this article was inspired and influenced mainly by the traditions of discursive analysis (e.g. Condor, 2011), critical discursive analysis (e.g. Wodak & Krzyzanowski, 2008; Van Dijk, 1984) and rhetorical approach (e.g. Billig, 2012). Much of the texts, especially the older ones, deal with the tensions around expressing prejudice in a world and era where there is a broadly shared prejudice against prejudice or a “general cultural norm against ‘prejudice’” (Billig, 2012, p. 141) and a tendency to consider prejudiced attitudes to be objectionable on ethical grounds and irrational in nature.^[6] Researchers have focused on the micro-level of strategic self-monitoring, self-presentation (relying on Goffman’s work) or more generally on identity management issues in expressing prejudice but avoiding the stigma of being evaluated as prejudiced (Condor, 2000; Augustinos & Every, 2007) or even more generally on “careful negotiation and identity construction around the topic of prejudice” (Wetherell, 2012, p. 168). For example Van Dijk explored specific semantic, pragmatic and conversation strategies of “adequate self-expression, positive self-presentation and effective persuasion” (1984, p. 116) when formulating “ethnic opinions” (p.116). Recently, the writings have focused even more on the “social” side—not on the verbal acts of individual actors but rather on how prejudicing and stereotyping happens as a result of the joint discursive action of several speakers (e.g. Condor & Figgou, 2012) or even of speakers and hearers (as a result of implicit allusions on one side and understanding of contextual information and shared categorical associations, e.g. Durrheim et al., 2016).^[7]

Thus, when looking now (in 2018) at the older materials, I am mainly interested in how people speak ABOUT “stereotypes” and “prejudice”, and for instance how they assess their accuracy and whether they admit to expressing views that can be considered prejudiced. In connection with the last point I assume that there are many more negative stereotypes around, which did not come up in the research projects, because the participants did not want to “admit” to them, meaning they did not consider it socially desirable to share them with us, the researchers, in that interaction.^[8]

In the following part I will show, using specific extracts from the empirical material, examples of different approaches to stereotypes/prejudice, the way the research participants talk/write and possibly think about them. From the wide range of aspects that could be focused on, I shall look in this article, in the following order, at prejudice as a source of knowledge; stereotypes declared as shared within the ingroup and (dis)agreement with them; different discursive ways of dealing with personal experience that contradicts a shared stereotype; and at a declared change in a stereotypical attitude following personal experience.

Results

Extract 1: I don’t know of any prejudice about Slovaks

Extract 2: In contrast to what we say here

(...) In contrast to what we say here about Germans, these two girls were much more spontaneous and friendly than me at the time (answer to a single open question “what is your experience of Germans?” in a questionnaire on intergroup contact and attitudes comprised of closed questions except for this one, 2010, Czech statement about Germans, female participant, age unknown, statement no. 1432).

The participant, speaking about a student exchange some time ago during secondary school, does not explicitly say what they “say here about Germans”. Again, if it was in an interview, at least it would have been possible for a vigilant interviewer to ask for more details about what “they” say and who “they” are, and whether the speaker thought so before, too. But we do not have this fuller answer and can only infer—for example, from the context of the positively coloured statement about receiving a friendly welcome while on the exchange, we could assume that this thing that is generally said about the Germans is quite the opposite of the speaker’s experience of the two spontaneous and friendly young German girls. This situation repeated itself several times, and always when the participants did not specify the stereotype or prejudice referred to, from the context it was clear that they were negative.

Extract 3: Ireland is beautiful.

(…) I have heard that the Irish just drink and take drugs and that Ireland is ugly and it’s always raining, but I came here and they are friendly and Ireland is beautiful.(…) (extract from an interview about experiences in Ireland and the existence of a Slovak “community” there, speaking about the decision to go abroad, 2009, AZ, age 29, male participant, short-term Slovak migrant in Ireland)

The personal experiences required for such a change were often not one-off, but repeated and/or long-term.^[13] Personal experience of the members of another national group does not always improve relationships and lead to the stereotypes being abandoned (see also Allport’s famous conditions for positive inter-group contact influence, in Allport, 1954). In the empirical material there were several cases mentioned where this reportedly did not work (cf. Paolini et al., 2010). For example, in one reported story, following personal contact among Czech and German secondary school students that did not go well a new negative stereotype was created (about what Czech secondary school students are like) and the whole exchange program was stopped.

Concluding remarks

I think stereotypes and prejudice are both a fascinating research issue and a topic with real everyday consequences for all of us. Given my vague suspicion of measurement tools such as direct self-report questionnaires,^[14] I felt enlightened and inspired when I discovered the work of several scholars that can be mainly grouped under the discursive and critical approaches to stereotypes/prejudice,^[15] who were not interested in how many people in group A would tick negative categorical evaluations of the members of group B. They had found so much more to investigate and problematize!

In this article I wanted to apply what I saw them doing with their extracts to the empirical material we had collected over many years and from many projects. Using a qualitative analysis inspired by the discursive approaches allowed me to observe how participants explicitly talk (and possibly think) about stereotypes—for example, how often and how exactly they mention them, assess their accuracy, (dis)agree with them, explain changes in their own originally stereotypical opinions, explain logically two contradictory assessments of members of the same national group in one short statement. However, I did not look at the most “classic” tension points, at the way speakers mitigate or manage expressions that could be judged as prejudiced in order to avoid being judged as prejudiced themselves. I was more interested in what I could learn about stereotypes/prejudice from the viewpoint of the participants, so in this article I have not used extracts containing prejudice/stereotypes (that I the researcher would have to evaluate as such) but ones ABOUT prejudice/stereotypes. That coincides with the claim of Condor and colleagues (2012) that the research on lay understandings of prejudice is surprisingly sparse, and with Billig’s recommendation (2012) that the research should include what ordinary people understand by “prejudice”, given that the concept is so important in lay discourse.

However, I kept to the individual level, just as the majority of measuring approaches do. Condor and Figgou (2012) criticize methodological individualism,^[16] as the main tendency among all the different approaches to prejudice and suggest that an alternative could be to think of prejudice in terms of collaborative cognition. In this approach groups or networks, not individuals, are the units of analysis. They show the construction, expression and suppression of subtle or blatant prejudice in a different light: first, they show the workings of so called “social scaffolding”—the way a more skilled person helps a less skilled person, instructing him/her in and facilitating the production of a logical (in this case racist) narrative. Second, they focus on how the pejorative portrayal of Others can be the result of joint action, where the contributions of each person are contextually important to the contributions of other participants, in this case allowing escalation in the expression of negative opinions. Third, they provide examples of joint inhibition, where one individual relies on others instead of self-monitoring and regulating his own expression of prejudice. Thus the display of prejudice is regulated through the interaction of several people—either through correcting the use of prejudiced categories or by openly admonishing the prejudicial talk of some of them— and not in the individual’s mind.

Moreover, Durrheim (2012), when writing about implicit prejudice in interaction, demonstrates how “stereotypes are formulated in the context of social interaction and that they can take an implicit form in which the hearer must help to stereotype” (p.190). In the same spirit, Durrheim and colleagues (2016) present an identity performance model of prejudice that focuses attention not only on how the expression of prejudice is responsive to norms and audiences but also how it shapes those norms. They also show how contestation of the very definition of what can and cannot count as prejudice, can be used either to mobilize hatred against out-groups (if their negative opinion of them is presented not as prejudiced but as reasonable), and to cement or change identities and norms. Regarding the latter they give an interesting example of paedophile people attempting to cast themselves as a “minor-attracted sexual orientation group” and the negative attitudes towards them as prejudice, which would result in very tangible real-life changes, in law for example. Further they illustrate how accusations and denials of prejudice “help to preserve categories, meanings and boundaries” (p. 26) and how repression of prejudice “can be viewed as a collaborative identity performance” (p. 29) in which all participants avoid the potential shame associated with uttering or hearing prejudice. Thus denials as well as accusations often remain inexplicit, but still the “work of reproducing the racial order” (2016, p. 29) has been done. Their article is a persuasive plea for a new agenda in social research that would attempt to “understand how the very definition of ‘prejudice’ is jointly defined and negotiated and deployed in social interactions to achieve social and political outcomes” (p. 32).

I consider these recent constructionist and critical psychological approaches to be highly inspirational but realize they cannot be applied to my empirical material—which I had at first hoped to do in order to obtain a kind of higher level analysis. Of course, having the empirical material that would allow for qualitative analysis (e.g. interview transcripts) is here not enough. To be able to “shake-off” the individual focus and to pay attention to the social or interactional, I would need “interactional” material—transcripts of conversations, for example. Ideally if I am to approach real-life situations these should not be elicited conversations (at the researcher’s request or in answer to a direct question) but either “natural” ones (see my explanation above) or ones that do not at least primarily focus on the issue interesting to the researchers (cf. Condor & Figgou, 2012). This last approach is exactly what I was trying to apply in this article when I began looking at older empirical material with a new topic and new perspective in mind.

To be more precise and honest, I wanted to put together material from over a long period and from several research projects, look at it with a new focus and then show how it could be analysed on several levels, inspired by the “classic” and more recent discursive analytical writing. Then I realized just how far the newest approaches have moved on and how inadequate my empirical material is for that.

I still think that the original idea of looking at the topic of stereotypes/prejudice using empirical material in which participants mention it spontaneously and not when prompted is a good one. But as became immediately clear, this does not work if there was not opportunity for letting them elaborate on the issue once touched upon. There are several reasons for this—the impossibility of asking further more detailed questions in the case of open questions in a questionnaire; interviews having a different focus at the time they were conducted and researchers wanting to cover too many topics in a single encounter in the case of interviews, and I would now say even the inability to explore the unexpected “jewel” of new and interesting information.

Thus in this article I was only able to go as far as the collected empirical material allowed, but exploring these new approaches has given me some new ideas for research that will be more difficult to realize but that will hopefully prove more helpful in the struggle to understand prejudice/stereotypes in the social reality of everyday life.