this research vary greatly, there is a common tendency to use interviews as an important part of triangulated data collection, along with observa- tion, diaries ...
RESEARCH ISSUES TESOL Quarterly publishes brief commentaries on aspects of qualitative and quantitative research. For this issue, we asked two researchers to discuss issues related to interview research in TESOL. Edited by PATRICIA A. DUFF University of British Columbia
Interview Research in TESOL Problematizing Interview Data: Voices in the Mind’s Machine? DAVID BLOCK Institute of Education, University of London London, United Kingdom ■
In recent years, there has been a noteworthy increase in the number of language education researchers publishing work that might be defined as ethnographically oriented. Although the questions being explored in this research vary greatly, there is a common tendency to use interviews as an important part of triangulated data collection, along with observation, diaries, letters, and questionnaires (e.g., recent TESOL Quarterly articles by Cox & Assis-Peterson, 1999; Flowerdew, 2000; Harklau, 2000; Ibrahim, 1999; Morita, 2000). When analyzing and discussing the data resulting from interviews, these and other researchers tend to focus on the content of the words produced by research participants, or as Freeman (1996) suggests, to take research participants “at their word.” The researchers relying on interviews for data are not oblivious to the problems this stance entails. However, given the space restrictions in journals and books, there is generally very little scope for discussion of this issue. What most readers encounter, then, is presentation of data plus content analysis, but no problematization of the data themselves or the respective roles of interviewers and interviewees.
TESOL QUARTERLY Vol. 34, No. 4, Winter 2000
757
PERCEPTIONS OF INTERVIEW DATA There is a long tradition in the social sciences whereby anthropologists, sociologists, sociolinguists, and educationists see interviews as conversations and co-constructed discourse events (e.g., Briggs, 1986; Burgess, 1984; Cicourel, 1964; Mishler, 1986) and therefore not as direct windows on the minds of interviewees. More recently, Kvale (1996) has contrasted two readings of interview data, the veridical and the symptomatic. In the former case, the accounts produced by research participants during research interviews are seen as reliable (or as reliable as possible) reports of events provided by well-meaning individuals. In the latter case, the analyst deems these accounts to be more about the research participant’s relationship to the topic and the interview context than about the topic being discussed. Thus when Kvale interviewed Danish secondary school students about grades, very often his participants provided accounts that proved to be veridically suspect, or in any case, contested (Kvale uses the term false) when compared and contrasted with other accounts by means of triangulated interviews. For example, students might state firmly that teachers gave better grades to students who spoke the most, but comments of teachers and fellow students might contradict this view. However, Kvale makes the point that the discovery of such inconsistencies did not invalidate the data; rather, it led him to see that the false (i.e., contested) accounts were symptomatic of his research participants’ feelings about grades rather than representative of phenomena that had occurred. Thus, students giving contested accounts of events were saying much more about their relationship with their teachers, or even about how they viewed the interviewer as someone to whom they could give negative information about their teachers, than they were about their experiences with grades. When I have analyzed interview data over the past decade or so, my first instinct has always been to take my research participants at their word, that is, to believe that they were providing veridical descriptions and evaluations of their lessons (Block, 1998). Despite this act of faith, I wonder if I might not conceptualize the data produced by interviewees in a different but complementary way, one that leans more towards the symptomatic end of the continuum described by Kvale. Following Freeman (1996), I might see interview data as representational of real events or as presentational of the individuals speaking. Adopting the latter stance means taking on some of the issues related to the presentation of self, such as how the interviewee constructs the interviewer, their relationship, and the purpose of the interview (see Block, 1995). In addition, I might move from seeing interview data as reflections of research participants’ memories of events (in other words, a cognitive phenomenon) to seeing them as reflections of how research participants 758
TESOL QUARTERLY
relate to the interview context as actors in a particular context (a social phenomenon). This means a rejection of the symbolicist perspective (i.e., underlying mental models that generate actions) and the acceptance of an interactionist perspective (i.e., interview data as a product of the interaction between interviewer and interviewee). Thus, interview data are not seen as the production of an individual interviewee but as the co-construction of interviewer and interviewee. These continua are presented in Figure 1. The conceptualization of interviews as co-constructions means that interview data are seen not as reflections of underlying memory but as voices adopted by research participants in response to the researcher’s prompts and questions. These voices might or might not truly represent what the research participant thinks or would choose to say in another context and on another occasion. However, at the same time, they must conform to what is allowed in what Gee (1996) calls a Discourse, defined as a socially accepted association among ways of using language, other symbolic expressions, and “artifacts,” of thinking, feeling, believing, valuing, and acting that can be used to identify oneself as a member of a socially meaningful group or “social network,” or to signal [that one is playing] a socially meaningful role. (p. 131)
Elsewhere, Lemke (1995) makes the point that we speak with the voices of our communities, and to the extent that we have individual voices, we fashion these out of the social voices already available to us, appropriating the words of others to speak a word of our own. (pp. 24–25)
In both Gee’s (1996) and Lemke’s (1995) ideas there is, of course, a link back to Bakhtin’s (1981) oft-cited concept of voice: the voices of others inhabit individuals’ voices, which in turn inhabit the voices of those with whom they participate in ongoing dialogue. What people say, therefore, is both constituted by and constitutive of the words of those with whom they share membership in a particular discourse community. FIGURE 1 Perceptions of Interview Data
➤ Representational ➤ Cognitive ➤ Symbolicist ➤ Individually constructed ➤ Veridical
RESEARCH ISSUES
➤ Symptomatic ➤ Presentational ➤ Social ➤ Interactive ➤ Co-constructed 759
Combining Gee’s, Lemke’s, and Bakhtin’s ideas, one could see any statement provided by a research participant as originating in a particular community that conforms to a particular Discourse and that may be classified as the kind of thing one is allowed to say in that community and within that Discourse. It is, therefore, data representative of that community, a particular voice that a particular research participant has adopted momentarily when providing an account. Elsewhere, Holstein and Gubrium (1995) add a further element to this conceptualization of interview data. They point out that during the course of an interview, interviewees might adopt different roles in response to their perception of being positioned in particular ways by particular questions. Thus in the research context described by these authors—individuals describing the type of home care they provide for their aging parents—interviewees responded in different ways depending on the stock of knowledge (Holstein & Gubrium, 1995) they drew on: Speaking as a daughter, the respondent is likely to frame her answer in terms of the events and sentiments of the daughter and mother’s interpersonal history. Speaking as spouse, the experiential purview of the respondent’s answers keys into relations with her husband and their domestic affairs. (p. 31)
The point is that in the course of an interview the same research participants might change voices depending on the way they situate themselves vis-à-vis a particular question and the person asking it. What is produced comes to be seen more as symptomatic of a particular state of mind and even ephemeral, ongoing social interaction than as a reflection of underlying memory or mental models of particular domains of knowledge and experience.
A CASE IN POINT To exemplify some of the points I have made here, I revisit an interview I carried out with an adult learner of English in Barcelona, Spain, several years ago as part of longitudinal study of adult EFL learners’ evaluations of their teachers and their classes. During the interview, I asked this learner (whom I shall call V ) questions about inconsistencies in class attendance, a phenomenon she had cited as a problem in earlier interviews. V responded to my questions, telling me about fluctuations in attendance, explaining how they played havoc with the rhythm of classes and, ultimately, adversely affected classes. Suddenly, however, in midsentence, her voice trailed off, she uttered a topic-closing “I don’t know,” and she said that she was ready to finish her English course. Noting her fatigue, I asked her if she was tired of studying 760
TESOL QUARTERLY
English, and V proceeded to tell me about a trip she had made to London during the previous week. She told me about her frustration at not being able to communicate in an effective manner with the Londoners she encountered, citing two episodes, one at a bar in a park and the other at a hotel. After reporting her problems understanding service personnel or being understood by them, she concluded by saying that she was ready to give up her studies of English. However, in V’s account of her London trip, I wondered if what she had produced was a veridical account or something more symptomatic. For example, when she went into detail about the frustrating experience in the park, she in essence contradicted herself. After saying, “When they talked to us it was like they were speaking Chinese!” and that the bar staff “didn’t really try to understand me,” she went on to say that she and her friends had not even tried to order and therefore had not exchanged words with the bar staff or anyone else present. One possible interpretation of this inconsistency is that V was simply making up her story as she went along. However, another way to see these data is as performances of plausible voices, in this case plausible learner accounts of interactions with English. In Holstein and Gubrium’s (1995) conceptualization, V draws on two distinct knowledge bases in the interview episodes of class attendance and her experiences in London. In the former case, she is giving voice to two widely held views among her language student peers about optimal conditions for language classes: Smaller groups are better than larger groups, and attendance should be consistent. Speaking as a tourist in London, V adopts the voice of the student of English who goes abroad and discovers, to her horror, that what she has learned in the classroom does not help her in an English-speaking environment. I do not wish to discredit V or call into question her honesty when providing her account of events; rather, her accounts of events in London may be taken as plausible stories, events that befall tourists and might have happened to her, but not necessarily exactly as she explained them. By the time this interview took place (late in her course), V was fed up with a situation in which variable attendance had led to the teacher’s losing control over the rhythm of the class, making class attendance less attractive and leading to more absences among students. Adopting the plausible voice of the inept tourist allowed her to express her frustration with her own progress in learning English and, more importantly, her discontent with her teacher, who ultimately was to blame for her lack of progress. Another consideration when examining this interview episode is V’s construction of me as her interlocutor. In other words, who was I in V’s mind? First, I was not V’s teacher, although I was a teacher in the school where she studied. Although V might have taken the position that anything she said might be relayed back to her teacher, she seems to have RESEARCH ISSUES
761
taken almost the opposite tack, using me as a depository for her expressions of frustration and discontent with her classes. More importantly, however, I think that V came to see me as a sympathetic ear, an individual who had nothing to do with the situation that was so frustrating to her but who understood the context and could speak to her about it in Catalan, her L1. In this sense, for V our meetings must have been more akin to therapy sessions than opportunities for information exchange. Indeed, about a week after V’s course and our last interview had concluded, she appeared at my office. When I told her that the study was effectively over but that we could still talk if she wished, she muttered something to the effect that it really did not matter and left. Thus, in some sense, it seems that the research interview was relaxed and unthreatening enough for V to want to come to talk about her classes each week and unload her frustrations related to her teacher, her classes, and the English language. In any case, it was more than just an information exchange between interviewer and interviewee.
CONCLUSION When researchers analyze interview data, they are not only studying representational accounts of events and the views of individuals; they are also confronting what is intelligible and plausible to say in a given discourse community and how members of that community use shared resources to construct a position in an interview. Interviews are therefore complex social and sociolinguistic events. Although adopting this broader view of interview data makes researchers’ interpretive task that much more difficult, it also allows them to see phenomena that might otherwise pass unnoticed were they to adopt a purely veridical view of their data (see Block, 1995, 1999). THE AUTHOR David Block works at the Institute of Education, University of London. His main interests are research methods, teacher education, second language acquisition (SLA), and culture in language teaching and learning. He is coeditor (with Debbie Cameron) of Globalization and Language Teaching (Routledge, in press) and is writing a book on the state of SLA.
REFERENCES Bakhtin, M. (1981). The dialogic imagination: Four essays. Austin: University of Texas Press. Block, D. (1995). Social constraints on interviews. Prospect, 10, 35–48. Block, D. (1998). Tale of a language learner. Language Teaching Research, 2, 148–176.
762
TESOL QUARTERLY
Block, D. (1999). Problematising interview data: What are they evidence of? Unpublished manuscript, Institute of Education, University of London. Briggs, C. (1986). Learning how to ask. Cambridge: Cambridge University Press. Burgess, R. (1984). In the field. London: Routledge. Cicourel, A. (1964). Method and measurement in sociology. New York: Free Press. Cox, M. I. P., & de Assis-Peterson, A. A. (1999). Critical pedagogy in ELT: Images of Brazilian teachers of English. TESOL Quarterly, 33, 433–452. Flowerdew, J. (2000). Discourse community, legitimate peripheral participation, and the nonnative-English-speaking scholar. TESOL Quarterly, 34, 127–150. Freeman, D. (1996). “To take them at their word”: Language data in the study of teachers’ knowledge. Harvard Educational Review, 66, 732–761. Gee, J. (1996). Social linguistics and literacies: Ideology in discourses (2nd ed.). London: Falmer Press. Harklau, L. (2000). From the “good kids” to the “worst”: Representations of English language learners across educational settings. TESOL Quarterly, 34, 35–67. Holstein, J. A., & Gubrium, J. F. (1995). The active interview. London: Sage. Ibrahim, A. (1999). Becoming black: Rap and hip-hop, race, gender, identity, and the politics of ESL learning. TESOL Quarterly, 33, 349–369. Kvale, S. (1996). InterViews. London: Sage. Lemke, J. (1995). Textual politics: Discourse and social dynamics. London: Taylor & Francis. Mishler, E. (1986). Research interviewing: Context and narrative. Cambridge, MA: Harvard University Press. Morita, N. (2000). Discourse socialization through oral classroom activities in a TESL graduate program. TESOL Quarterly, 34, 279–310.
Conducting Individual and Focus Group Interviews in Research in Albania SILVANA DUSHKU University of Illinois, Urbana-Champaign Urbana, Illinois, United States ■
Interviewing is one of the most commonly used survey methods in present-day qualitative research (Berg, 1998; Denzin & Lincoln, 1998; Jones, 1985; Kvale, 1996; Mishler, 1986; Morgan, 1993). Used to elicit qualitative data through a social interaction between the interviewer and the interviewee(s), interviewing can provide information on reported behavior, attitudes, and beliefs, and contribute to an in-depth understanding of research participants’ perspectives or experiences (Walker, 1985). However, interview research in developing countries, such as Albania, can be complicated by economic, political, and social constraints on local contexts and on participants that result in increased emigration, turnover, workloads, and overall instability. In this report, I describe a study I conducted in Albania in 1995–1999 to evaluate English
RESEARCH ISSUES
763
