• Tidak ada hasil yang ditemukan

Excerpt 6 June 1988)

4.10 Interviews

Interviews play an important role in much research in sociolinguistics and applied linguistics/SLA. Not surprisingly, they are very commonly used in case studies (see Section 4.8; Merriam, 1998). Many helpful books and articles deal with research interviewing in the social sciences and humani- ties (Briggs, 1986; Fontana & Frey, 1994; Holstein & Gubrium, 1997; Kvale, 1996; Miller & Glassner, 1997; Mishler, 1986; Seidman, 2006; Spradley, 1979; Schostak, 2006; Yow, 1994). In applied linguistics, some recent text- books also devote considerable space to practical approaches to qualitative interviewing (e.g., Chapter 2 in Richards, 2003). In this section, I summarize some of the main points of relevance in qualitative case study research.

There are various ways of categorizing interviews. For example:

Structured, semistructured, or unstructured interviews In-depth interviews (either semistructured or unstructured) Focus group or group interviews

Oral history or life history interviews

The amount of control or structure imposed by researchers on the interview format and resulting speech event varies across studies. Many early—(post)positivist—case studies of L2 learners (e.g., Hatch, 1978a) involved interviews eliciting L2 production through the use of pictures, conversational question-answer segments, personal narratives, and think-

aloud protocols in what was construed to be an oral interview (i.e., as dis- tinct from a test or lab experiment). The goal was to learn more about the full extent of language learners’ (oral) competence/performance in their first, second, or additional languages. The use of a variety of tasks helped elicit as many types of discourse or targeted structures and as much ana- lyzable language as possible to understand learners’ true abilities and sys- tematic language use.

A second approach to linguistic interviewing draws more on Labov’s (1966) classic work using personal, sometimes affectively charged, nar- ratives of past events primarily, particularly for studies of phonological variation, tense-aspect systems, narrative structure, and discourse-level phenomena in casual, everyday speech. Again, the focus is primarily structural (formal or functional).

A third approach uses interviews primarily to collect data about the insights or perspectives of research participants, with less attention paid to the actual linguistic or textual features of the discourse. A content or thema- tic analysis, rather than a linguistic or interactional analysis, is primary.

Phenomenological research is another approach that focuses on peo- ple’s “lived experiences” with respect to language, culture, education, immigration, and so on, in the humanities, social sciences, and education.

This approach places greater emphasis on participants’ narrative recon- structions of aspects of their lives and experiences, such as the connec- tion between their affective or emotional states or their identities and their experiences of language learning (e.g., Kouritzen, 1999; Pavlenko, 2002;

Schumann, 1997). In content- and narrative-oriented research such as this, it is common to conduct more than one interview with interviewees. In that way, the researcher can follow up on issues or clarify uncertainties emerging from an earlier interview. The purpose is not usually to examine developmental changes in this instance, but to seek clarity and perhaps consistency in accounts as well as an elaboration on significant topics. Life history interviews fall into this category.

Whichever approach to interviewing is taken (and there are others), it is important to recognize that a research interview is a “construction” or joint production by interviewer and interviewee (Briggs, 1986; Coughlan

& Duff, 1994; Mishler, 1986). It produces a version of truth, a snapshot of competence or of ideas elicited for a specific purpose in a particular space and time. The data are generated by means of social interaction between

interviewer and interviewee and cannot necessarily be taken as decon- textualized, independent facts or observations. Each interview has, and is, its own discourse context, which also evolves over the course of the interview and from one interview to the next. Even research on oral pro- ficiency interviews in language testing in recent years now acknowledges the highly interactional and contingent nature of interviews (e.g., Young &

He, 1998).

As a type of speech event, moreover, interviews may have a differ- ent cultural status or discursive form in different societies. In addition to being interactive, cultural, and dynamic, there is often a power differential between interviewer and interviewee as well. In Learning How to Ask, Briggs (1986) famously reflected upon his own communicative problems (“blunders”) doing ethnographic research interviews with Mexicano Span- ish speakers in New Mexico, and also in trying to make sense of interview data. He critiques interviewing in social science research on a number of levels, one of which is sociolinguistic:

By participating in an interview, both parties are implicitly agreeing to abide by certain communicative norms. The interview moves the roles that each normally occupies in life into the background and structures the encounter with respect to the roles of interviewer and interviewee.…

Some potential respondents are drawn from communities whose socio- linguistic norms stand in opposition to those embedded in the interview.

(pp. 2–3)

Just as some interviewers are better or more adept at interviewing than others, some respondents (interviewees) are better in that role than others.

They are more forthcoming, reflective, analytical, proficient, articulate, vivid, or even more interested in the research itself than others might be.

The chemistry between the interviewer and interviewee (in addition to context, purpose, history, etc.) may have something to do with why some respondents provide better data. In selecting cases for study (see Section 4.5), it is important to anticipate how well participants will perform in interviews, if they are to be used, and how as an interviewer one might deal with case study participants who are reticent, uncooperative, inarticu- late, verbose, or easily sidetracked. Choosing the most articulate candidate interviewees, however, may introduce some skewing of data because they

may not be very representative of others in their category. They may have had the most successful experiences learning language or becoming accul- turated, for example, or may be more confident or more willing to take risks in unfamiliar social contexts than their peers.

Qualitative research interviews are normally conducted face to face, especially with L2 users, but if it is difficult to arrange meeting times or places, the telephone may provide a useful—albeit imperfect and some- times challenging—substitute. Audio equipment can be easily found to record the talk, if permission has been granted to do so. Increasingly, e- mail exchanges or threaded discussions can also supplement or take the place of formal interviews, depending on their purpose.

Although one-on-one interviews between researchers and research par- ticipants are probably most common in applied linguistics and especially in case studies, in multiple-case studies they could be supplemented with focus group interviews. In focus groups, groups of research participants, for example, six to eight per group, are asked to discuss their perspectives on issues. While perhaps better known in the fields of marketing, soci- ology, public health, political science, and women’s studies, focus group interviews can reveal in a fairly short time frame several people’s perspec- tives on an issue, sometimes referred to as multivocality (Duff, Wong, &

Early, 2000; Krueger, 1994; Madriz, 2000). The group interaction itself can prompt others to comment on themes that they might not have thought of or volunteered in one-on-one settings, and participants often find the group format less intimidating than a one-on-one interview. At the same time, sensitive topics are best not broached in such a forum in order to pro- tect the privacy, anonymity, and comfort level of research participants.

Focus group data can be useful either for interpreting case study data against the backdrop of a wider pool of subjects or as a strategy allowing researchers to carefully select possible cases for subsequent study. It is a less effective way of collecting data for the analysis of people’s linguistic competence (unless interaction during a focus group interview is itself the topic), and it can be difficult to identify speakers during transcription using audiotape recordings alone unless the focus group participants’ voices are familiar and do not overlap too much.

Some practical tips for interviewing either individually or in groups are found in Table 4.3.

The kinds of questions that might then be asked in an interview are found in Table 4.4.

Ideally, interviewers themselves should demonstrate as many of the attributes found in Table 4.5 as possible (Kvale, 1996).

Normally, interviews are audio-recorded with suitable high-quality tape recorders and microphones in a quiet location. Increasingly, digital

Table 4.3  Practical Tips for Interviewing

1. Carefully craft the questions to be asked, in terms of both content and form, based on the main focus of the study, and think of the optimal time and location for the interview.

2. Eliminate unnecessary jargon from questions. For example, in seeking insights from immigrant students or workers about their identities, the questions “What is your main ethnolinguistic identity?” or “What are the various identities you negotiate in your daily life?”

may draw blank stares (as they have in some UBC research). It is better to paraphrase the concept or exemplify it in terms that people can easily understand.

3. Ask a colleague or nonspecialist to review the questions before the interview or try them out (pilot test) with a nonparticipant to ensure that the language is clear, the questions are not too complicated, and the interview is not too long for the allotted time.

4. Check your recording equipment carefully before each use. Ensure that batteries are fresh, that cassettes (if used) are available and carefully labeled, and that microphones and power supplies are prepared, as needed. Consider bringing and using a backup recording instrument for high-stakes interviews in case the main machine malfunctions. Find a quiet venue for recording.

5. Start the interview with some small talk to put interviewees at ease. Then follow the interview protocol (or set of questions) with whatever strictness or flexibility is necessary or appropriate.

Table 4.4  Types of Questions in Interviews

Introducing questions (e.g., “Can you tell me about your L2 learning history?”)

Follow-up questions (asking more about the previous utterance or response) e.g., “Could you say more about that?” or “Could you give an example of that?”

Probing questions (e.g., “Could you say something more about the courses you attended?”)

Specifying questions (e.g., “How did you react?”)

Direct questions (e.g., “Have you ever had trouble being understood?”)

Indirect questions (e.g., “How do you think other students feel about x?”)

Structuring questions (moving onto another topic when off-track or when topic has been covered; e.g., “Now I have a different question/topic/task for you.”)

Silence (to encourage reflection or amplification of responses)

Interpreting questions (e.g., “You mean that …?”)

Source: Adapted from Kvale, S., InterViews: An Introduction to Qualitative Research Interviewing. Sage, Thousand Oaks, CA, 1996, pp. 133–135. Adapted with the permission of Sage Publications.

recorders are used because the data can easily be uploaded to computers for replay, transcription, analysis, and presentation. The existence of too much background noise will make transcription much more time consum- ing (and expensive) and onerous; noise is also distracting during an inter- view. Semistructured or unstructured interviews usually last anywhere between 30 and 90 minutes.

The language of interviews (or use of interpreters) and the gender of interviewers in relation to their interviewees may be important consider- ations in research with culturally and linguistically diverse populations (Goldstein, 1997; Kouritzen, 1999). In some contexts, interviewees might be more forthcoming with an interviewer of the same sex who shares the same L1 and culture. However, in other cases, interviewees might be reluc- tant to participate in interviews with people from the same backgrounds for fear of judgment or gossip from within that community, or they might find it easier to speak the L2 (e.g., English) with interlocutors from other linguistic groups if their L2 oral proficiency is being analyzed because it would be more natural. If interviewers are closer to research participants in age and status (for example, graduate students interviewing other students), it is often easier to develop rapport and to put the research participant at ease (e.g., Kobayashi, 2003; Morita, 2004). For this reason, in several of my recent studies at the postsecondary level, I have hired research assis- tants who come from the same linguistic backgrounds as the students they

Table 4.5  Desirable Interviewer Attributes

Knowledgeable, respectful, organized, and clear, both in language use and purpose for the interview

Aware of how much they themselves are speaking, and for what purpose

Careful not to ask leading questions that will suggest a certain desired or preferred response or perspective

Attuned to the level of proficiency or comprehension of the interviewee

Careful not to interrupt and trained to provide sufficient response time following questions

Sensitive, attentive, and responsive to answers or questions (in less structured interviews)

Open to relevant new directions (when appropriate) Able to keep discussion on track, but do so gently

Able to remember what interviewees have already said Effective in interpreting interviewees’ remarks

are interviewing. Their in-group status as peers at the same institution has greatly facilitated the data collection process.

Finally, in ethical review applications and in many published reports, the researcher often needs to provide interview guides or protocols in an appendix with sample questions to be asked for reference purposes.

Because gaining access to fieldwork sites for direct observation is becom- ing increasingly complicated for ethical/institutional reasons in some types of research (see Section 4.14), case studies in applied linguistics will probably rely on interviewing and written text analyses to an even greater extent as their principal data collection strategies in the future (see Duff, Wong, & Early, 2000).

Dalam dokumen Case Study Research in Applied Linguistics (Halaman 143-149)