‘Some of our people can be the most difficult’. Reflections on difficult interviews.
by Sam Pryke
Liverpool Hope University
Sociological Research Online, 9 (1) pryke
Received: 22 Jul 2003 Accepted: 20 Feb 2004 Published: 28 Feb 2004
This article questions the methodological convention in the social sciences that the interviewer must never disagree with a respondent in qualitative research. The issue arose during research on the British Serbian community when some participants sought to justify, exculpate or reject Serbian liability for atrocity. My initial response not to demur but to simply move onto the next question morally tainted the research, as it seemed to collud in a denial of Serbian responsibility for atrocity in an understanding of war (1991-99) in which the Serbs were always the victim. I discuss, through an extended excerpt from an interview conducted later in the research, my attempt to challenge respondents over this claim. I set the moral and methodological case to object to the denial of atrocity against the practical dangers present in doing so: the risk of a loosing track of the spine of a prepared script of questions as a fruitless argument develops and the intricacies of the subject matter are exposed. But I also allow for an interpretation that would suggest that my response was altogether too cautious. My conclusion, such as one can make one about such a complex matter, is that to object in such a kind of instance is legitimate.
Keywords: Serbs, interview, atrocity, denial, argument, ethics.
I just find it amazing in Bosnia, it goes back to that joke tape that we have got, you know the Serbs claimed all that land, they were massacring, they were raping, where the hell did they find time to do everything?! They weren't the numbers there, it wasn't as if the Serbs outnumbered the Muslims, but where did they find the time to rape 30,000 women in all that time, do all the massacring and secure and gain all that territory?
Disagreement in Qualitative Research
Researching the British Serbian Community
Talking about War
Interviewer: At that time, 1991/2, did you support Milosevic, despite his Communist Party past, as a figure who was championing the cause of Serbs throughout Yugoslavia and did you support the idea of a Greater Serbia, understood as you say as a defensive reaction?
Respondent one: The simple answer is yes. At that time the various ethnic groups in that region wanted their own independence. The difference between the Serbs and the other groups is that a significant minority of the Serbs at that time lived outside of Serbia. The only way that those Serbs could feel safe was in a state of their own, whether or not you call it the state of Yugoslavia or Greater Serbia.
Respondent two: They didn't want a Greater Serbia for the sake of a Greater Serbia to make a kind of superpower, but they had to be together. So yes, I supported that and I think most people here did. Did you support Milosevic as a figure who represented that?
Respondent one with general agreement: Yes, he gave that impression at the time.
Was that true throughout the community? What was the view of the first generation?
Respondent one: Well there were a few of the first generation who didn't support him at first, but they came round. It got to a point where people, regardless of whether he was a communist or a royalist, saw that he was doing it for Serbs, as long as he was doing it for Serbs. And there was a great feeling that for the first time we were going to be united. Initially, that was what he wanted to do but the West got involved.
Respondent two: The UN, the Western World, had this idea to try and impose a new world order on Yugoslavia.
Well whatever outside powers did they weren't responsible for atrocities on the ground. The evidence would suggest that in respect to Bosnia, the worst atrocities were committed by the Bosnian Serbs because of their control of the Yugoslav army. Did you at any point think that the creation of a Greater Serbia would inevitably involve the mass expulsion of Croats and Muslims?
Respondent three: Perhaps I am naive but I didn't think that it would.
Respondent one: I don't agree with the statement you just made, but it was a bi-product of the war.
Respondent two: A lot of it was voluntary, people just went to live with people of their own background and culture. And some people stayed. I just find it amazing in Bosnia, it goes back to that joke tape that we have got, you know the Serbs claimed all that land, they were massacring, they were raping, where the hell did they find time to do everything?! There weren't the numbers there, it wasn't as if the Serbs outnumbered the Muslims, but where did they find the time to rape 30,000 women in all that time, do all the massacring and secure and gain all that territory?
Are you saying that the rapes and murders didn't happen?
Respondent one: Yes they did, on all sides. They happened, they happened, we are not saying they didn't.
Was there ever a point when you felt you had to condemn it?
Respondent one: But it goes back to the victors truth. America and Great Britain are not going to admit that they cocked up, history will prove that they cocked up. But they are not going to admit that.
Well, I think that some external figures who got involved did concede that they got things wrong. But my own view is that although outside powers may have made things worse, there's a case for saying that, the responsibility of what happened there was that of the people who lived in that part of the world. But I just wondered, how long your support for the Serbian cause remained, perhaps it never waivered, or were there points when you thought it necessary, perhaps only to yourself, to condemn what happened there?
Respondent two: It made me feel what was going on out there, my own view, I felt more proud to be a Serb. I felt the longer it was going on, the more misinformation we were given, it made my feelings and my own proudness to be a Serb even stronger. Because you only had to look at the bombing of the bread queue in Sarejevo: front pages, headlines 'Serbs bomb bread queue'. It was proven by the UN that it was a Muslim attack on their own people for their own propaganda purposes.
So your support for the Serbian cause didn't waiver at any time?
Respondent two: No it just got stronger and stronger.
And that was furthered because of the portrayal of Serbs as the aggressors?
Respondent one: It became more and more evident that the media, the governments, they'd all taken sides, they'd decided that the Serbs were the bad guys, the others were the good guys. And they did their best to present those parties in that light. Any evidence they could use against the Serbs they did. Any evidence to support the Serbs they ignored.
1) The most general point is of a theoretical/politcal: explicit disagreement represents a serious deviation from the role of the interviewer in sociological research. It might be argued that a respondent should not seek to contradict a respondent even if disagreement is couched in relatively mild terms like ‘the evidence would suggest’, ‘in my view’. As indicated above, there is recognition that completely neutral questions are impossible, that it is quite legitimate to probe a respondent’s answers and permissible in some circumstances to formulate ‘devils advocate’ questions. But this is some way from making pre-emptive claims – in a style one critic compared to a Hague Tribunal prosecutor - in an interview. The sociologist is an interpreter not a legislator. If he/she has pretensions to change minds then at the very least this should be somehow acknowledged at the outset of the interview, and thus placed in the context of some form of action research or, to extend Bourdieu, ‘sociotherapy’. But if a researcher has such moral scruples that make it impossible for him/her to encounter views that they find objectionable, they should not undertake the research in the first place.
2) It is questionable what such questioning actually achieved as, other than conferring a certain moral and, possibly masculine self-righteousness – restated here in retrospect -, rather more information might have been attained by a less confrontational strategy based upon maintaining a positive rapport.
3) To seek to embark on such an interviewing strategy necessarily places the interviewer in a position of spurious authority whereby he/she assumes that they have a controlling and necessarily superior knowledge about the subject matter than the respondent. 4) This line of questioning inevitably raises the issue of whether or not it is legitimate to try to bring an alternative morality with people who, through in this instance an overriding national allegiance, see things rather differently.
5) To contradict in such a context risks entering a conversation thicket from which it is difficult to extricate oneself from, whilst maintaining any degree of control over the direction of the conversation. Therefore the strategy both assumes omnipotence and jeopodises elemental control.
5.6 I have no wish to try to answer all such possible criticisms, especially as I think aspects of some valid. They were composed with the benefit of hindsight and with the benefit of hindsight I would have done things rather differently. Who in retrospect would not have conducted primary research differently? This said, I would wish to defend and indeed advocate the interviewing strategy used above. For the sake of clarity I will outline my thoughts on this as responses to the five points made above.
5.7 First, whilst there is an absence of direct reference to disagreement with respondents within interviews in both theoretical and practical discussion of sociological research, the overview of the issue at the beginning of the article indicates that there is sufficient flexibility within prominent approaches to incorporate it in my research. Although it is the case that my decision to try and challenge certain opinions was not directly prompted by a prior theoretical position, I was influenced by Bourdieu’s concept of socioanalysis and was interested to see how it would ‘play out’ within interviews. Moreover, the interviewing strategies recommended by Minchiello (1991), were of some importance. The possible criticism that sociology is not directly concerned with changing minds through argument within interviews, seems based on an assumption that opinions are static. Several of the people I spoke to – though not those quoted above – told me that they had come to think about things slightly differently as a result of being interviewed. Now clearly to have set out to proselytize would have been at once arrogant and naive. But my determination to directly disagree with respondents was not motivated by a fanciful ambition to re-educate Serbs. Rather, I did so because to continue to have remained quiet would have been morally wrong and methodologically loose. Having said this, I think that it would have been advisable if I had stated candidly before asking questions about the break-up of Yugoslavia that, whilst I was not setting out to disagree with respondents, I might well venture another opinion. This would have added to a certain ethical transparency to the dialogue.
5.8 Second, I certainly think it was the case that being prepared to disagree with the respondents made the interview far more revealing than it otherwise would have been, as it established at least four things. First, clarification that the respondents supported the attempted creation of a Greater Serbia, albeit legitimated as a defensive strategy. This is important as the idea that there ever was a definite attempt to create a Greater Serbia, as opposed to defending Yugoslavia, is sometimes glossed over or even denied within Serbian circles. Second, in contrast to the tack I had taken in some previous interviews, it pressed respondents on whether this inevitably involved ethnic cleansing as a logical consequence of this. Third, it repeatedly pushed the issue of Serbian responsibility for war by my refusal to accept the facile contention that it was all the fault of external powers. Fourth, it established that several of the respondents acknowledged that Serbs did commit atrocity, specifically rape, after the issue had initially been brushed aside through sarcastic humour.
5.9 So, disagreement was not just an act of moral self-dramatisation, but something that gave greater depth to the data obtained. I am doubtful that an less critical strategy that tried to keep alive a positive rapport at all costs would have been equally as effective. Moreover, as indicated, it is the case that a critical dialogue has a certain integrity as compared to a silky rapport that lures a respondent into disclosure with no hint prior to the write up that the interviewer finds their testimony objectionable. This is not to suggest, of course, that a critical dialogue precludes moments of rapport (Campbell 2003: 295).
5.10 The third possible objection that such a strategy assumes a position of unwarranted authority in which the interviewer stakes out ‘the truth’ within the interview, involves interpersonal and epistemological issues. Anyone who has done qualitative research will know that power relations within interviews, as in interaction more generally, are always subject to some degree of ongoing negotiation or ‘dialectic of control’ (Campbell 2003: 297). In the above interview, it was the case that the respondents seemed generally strong minded and forthright individuals who were never likely to defer to the opinion of an academic – who does these days? Prior and subsequent conversation revealed that they knew a good deal of Balkan history, spoke and read Serbian well and visited relatives there fairly frequently. In certain respects their knowledge and expertise were greater than mine. Moreover, they had the ‘personal capital’ of talking about ‘their country’ – they had lost relatives in August 1995 and the land their family had inhabited for centuries. Finally, they were probably conscious of their position as gate keepers as they knew that my research depended on an ongoing chain of contacts. So, far from the interview being an encounter where I had utilised a controlling academic knowledge and authority, I felt somewhat battered and emotionally drained afterwards.
5.11 The issue as to whether an interviewer can set out the truth as to what happened in an issue like a civil war is in some ways more complex. Take the above example of the Sebrenica massacre that I did not pursue in an early interview. Extended questioning about this episode would not have been quite as straightforward as might appear. When told that the CIA were involved in the provocations of Serbs around Srebrenica, I assumed that the claim was just another Serbian conspiracy theory. However, a BBC ‘From Our Correspondent’ BBC documentary, screened in January 2001, revealed this to be essentially accurate. This scarcely justifies the slaughter of up to 8,000 men and boys in cold blood. The point, however, is that almost inevitably in a bloody civil war that pitted neighbour against neighbour, no party was without blame. In fact, all parties tended to claim that any particular act was motivated by a prior attack – ‘they did it to us first’, ‘we only did to them what they did to us’, being the standard rationalisations -, and in the wider scheme of things they were the greater victim. My basic point here is that an interviewing strategy that involved contradicting respondents and stating an alternative view is somewhat more problematic than might at first be thought as no final and complete history will ever be told of the break-up of Yugoslavia, or indeed any other historical event of such complexity. There was obviously not therefore a touchstone of truth that might some how be called upon. Nor should a research interview by conceived as akin to a courtroom trial in which evidence is displayed and scrutinised in the interests of determining the truth. It concerns the experience and opinions of a research subject. However, this important acknowledgement is quite different from a position in which truth becomes a causality of rival accounts of events or of relativist postmodern philosophy in which the interview is only of interest as a self-referential text (Alversson and Sköldberg 2000: 4). We know, for instance, that rape was carried out on a large scale by Serbian forces, especially in Bosnia 92-93, not just as a spoil of war but as an aspect of systematic ethnic cleansing. We know this because of the painstaking work by, in particular, feminist researchers who have interviewed survivors and uncovered Serbian documents on the taking of territory (Stiglamayer 1994). If in an interview a respondent dismisses the idea that mass rape really happened at all, it seems therefore legitimate to pursue a line of questioning that seeks that takes issue with him or her.
5.12 The fourth possible criticism, one of possible opposed moralities, has no foundation. On the one hand, it would rely upon a conception of ethics that delineates sociological research from everyday life (for critique seeStanley and Wise 1993: 203). If I heard somebody deny that mass rape ever happened despite overwhelming evidence that it did, I would almost certainly object. Even allowing for the some greater care in the choice of words used, why should a research interview be different? Moreover, a process of denial does not hinge upon an alternative morality. On the contrary, it accepts a shared morality. It was not as if the respondents claimed that, yes, Muslim women were raped and that this was a legitimate means of violence and impregnation in ethnic war. Rather, at least one of the respondents initially made light of the suggestion that it had ever happened at all.
5.13 The fifth possible problem is that once an interviewer begins to disagree with a respondent even in measured terms a polemical momentum is quickly established. The interview quoted from above did not culminate in shouting and abuse followed by blows, but looking over the transcript I can see that that my subsequent questions were not so carefully phrased and, in fact, became more like statements and closed interrogative points, habitually used in the cut and thrust of argument. So there is a danger involved of a row developing. However, the interview like the research process in general is beset by potential problems and risks (Stanley and Wise 1993: 153). If one was intent upon avoiding them altogether it is unlikely that much sociological research would ever get done.
Taking disagreement a step further?6.1 There is of course a quite different response to the interview quoted above: my response did not go far enough. One could reasonably argue that the kind of claims constituted an incoherent but definite process of denial: the implausibility of claims of atrocity, the moral equivalence of war, the primary responsibility of external powers, the demonisation of one side by the media, and, incredibly, the claim that a process like ethnic cleansing was voluntary. Nobody could seriously argue that a directly confrontational argument within an interview would benefit anybody and serve to take a research project forward, but one might argue that it is the responsibility of the sociologist to attempt in some way to try to break through denial by offering counter interpretations, rather than just simply making the questions somewhat more critical.
6.2 For instance, it became apparent to me in the course of the research how a single incident, the market place bombing in Sarejevo, 5th February, 1995, was generalised. In his otherwise thorough book States of Denial (2001), Stan Cohen does not consider how a particular event can become a template for a wider counter-interpretation. The claim, by Serbian respondents as evidenced above, was that contrary to the initial media assumptions that the Serbs had fired the mortars, it was officially proved by UNPROFOR that the Bosnian army had actually attacked their own people in order to further discredit the Serbs. In fact, this was not the case as UNPRFOR ballistics experts were unable to confirm with any certainty the location from which the missiles had been launched. This made for controversy, although there remains some doubt as to whether this uncertainty was real or simply a form of leverage used by the UNPRFOR commander General Michael Rose to reverse the Bosnian government refusal to take part in negotiations over demilitarisation of the city, scheduled for February 8th (Allcock, Milivojevic and Horton 1998: 259).
6.3 Notwithstanding the lack of clarity here, I might have made this specific point in interviews, rather than simply passing on without dissent to the next question or stating a general disagreement. Or I might have taken things further still. The extraordinary claim, ‘It made me feel what was going on out there, my own view, I felt more proud to be a Serb’, might have been countered with something like, ‘With respect, how can anyone be proud of the siege of Dubrovnik, or the destruction of Vukovar or the massacre or Sebrenica in a war in which your side, the Serbs, were in any case the ultimate loser?’. The strategy would have risked confrontation, but might have been more revealing and transformative.
6.4 This strategy would, I think, have been highly problematic. It is altogether more likely to turn the dialogue into a time consuming and fruitless row which generates more heat than light, as both parties lose all track of the general trajectory of the interview. A withering set of questions about a single event, or a wider refutation of an objectionable opinion when the interviewer knows the respondent is wrong, might appear attractive to a sociologist who admires how a combative TV journalist attempts to prize information from a government minister. However, the two types of interview are quite different and the means of eliciting knowledge cannot be directly compared.
Conclusion7.1 This article has sought to make the case that it is permissible to directly disagree with respondents in an interview in the light of my experience in conducting research on the British Serbian community. Important trends in thinking about methodology would seem to allow for the possibility of disagreement, but it is notable that there is no direct discussion of the matter in either the theoretical or more practical literature on qualitative research. My determination to risk argument within interviews was not wrought by a vain attempt to inaugurate a new methodological departure, but the experience of initially listening without dissent to the attempts by second generation British Serbs to exculpate Serbian actions during the break-up of Yugoslavia, 1991-1990. In retrospect I felt that my silence had colluded with the denial of atrocity and, relatedly, allowed the assumptions that underpinned the accounts to go unchallenged. As a result, I subsequently contradicted respondents within interviews by being prepared set out what has been established about the war. Although it did not necessarily make the interview quoted above any more pleasant at the time, the strategy did succeed in uncovering respondents’ beliefs in a way that a series of less disputatious questions would not have done. Such a research strategy has, as I acknowledge, inherent dangers. However, in certain circumstances the risks are worth taking given the both moral and procedural issues at stake. A strategy of openly disagreeing with respondents should therefore be added to the existing stock of research techniques in qualitative research.
ReferencesALLCOCK, J.B., MILIVOJEVIC, M. and HORTON, J. J. (editors) (1998) Conflict in the Former Yugoslavia, An Encyclopedia. Denver, Col.: ABC-CLIO. ALVESSON, M and SKÖLDERBERG, K. (2000) Reflexive Methodology, New Vistas for Qualitative Research.London: Sage.
BELL, C. and ROBERTS, H. (editors) (1984) Social Researching: politics, problems and practice. London: Routledge.
BHABHA, H. (editor) (1990) Nation and Narration. London and New York: Routledge.
BOURDIEU, P. et al (2000) The Weight of the World: Social Suffering in Contemporary Society. Cambridge: Polity.
BOURDIEU, P. and EAGLETON, T. (1992) ‘A conversation about doxa’, New Left Review, No.191.
BULMER, M. (editor) (1982) Social Research Ethics: an examination of the merits of covert participant observation. London: Macmillan.
CAMPBELL, E. (October-December 2003) ‘Interviewing men in uniform: a feminist approach?’, Social Research Methodology Vol 6 No 4., pp. 285-305.
COHEN, S. (2001) States of Denial, Knowing About Atrocities and Suffering. Cambridge: Polity.
DAVIES, A. C. (1999) Reflexive Ethnography, A guide to researching selves and others.London: Routledge.
DENZIN, N. AND LINCOLN, Y (editors) (2003) Collecting and Interpreting Qualitative Materials. London: Sage.
FINCH, J. (1984) ‘“Its great to have somebody to talk to”: the ethics and politics of interviewing women’ in C. Bell and H. Roberts (editors) Social Researching: politics, problems and practice. London: Routledge.
FONTANA, A. and FREY, J. (2003) ‘The Interview: From Structured Questions to Negotiated Text’ in Denzin and Lincoln (editors), Collecting and Interpreting Qualitative Materials.
HOLSTEIN, J.A. and GUBRIUM, J.F. ‘Active Interviewing’ in D. Silverman (1997) Qualitative Research, Theory, Method and Practice.London: Sage.
HAMMERSLEY, M. (editor) (1993) Social Research, philosophy, politics and practice. London: Sage.
MAGAS, B. (1993) The Destruction of Yugoslavia, Tracing the Break-up 1980-1992. London: Verso.
MINCHIELLO, V., ARONI, R., TIMEWELL, E. and ALEXANDER, L. (1990) In-Depth Interviewing. Melbourne: Longman Cheshire.
MCKENNA, Y. (2003) ‘Sisterhood? Exploring power relations in the collection of oral history’, Oral History, Vol. 31. No.1. pp. 65-72. PADGETT, D.(1989) Settlers and Sojourners: a study of Serbian adaptation in Milwaukee, Wisconsin. New York: AMS Press.
PROCTOR, N. (2000) Serbian Australians in the Shadow of the Balkan War.Aldershot: Ashgate.
RENAN, E. (1889) (1990). ‘What is a nation?’ in H. Bhabha (editor) Nation and Narration. London and New York: Routledge.
SILVERMAN, D. (editor) (1997) Qualitative Research, Theory, Method and Practice.London: Sage.
STANLEY, L. and WISE, S. (1993) Breaking Out Again, Feminist Ontology and Epistemology.London and New York: Routledge.
STIGLMAYER, A. (editor) (1994) Mass Rape against women in Bosnia-Herzegovina. Lincoln: Lincoln University Press.
TODOROVA, M. (1997) Imagining the Balkans. Oxford & New York: Oxford University Press.
TURNBULL, A. (January-March 2000) ‘Collaboration and censorship in the oral history interview’, Social Research Methodology, Vol 3 No 1., pp. 15-35.
WENGRAF, T. (2001) Qualitative Research Interviewing, Biographic, Narrative and Semi-Structured Methods. London: Sage.