June 24, 2007

Interview methods research: A thin line between qualitative and quantitative family research methods

What is the responsibility of quantitative family researchers to be familiar with and contribute to interviewing methods research? On the one hand, complex interviewing methods are nothing new to quantitative family researchers: writing useful interview questions, choosing interviewing techniques and formats, recruiting and training interviewers, considering the effect of technology choices on participants and coders, creating coding manuals, training coders, and maintaining coder reliability are just a few methods issues family quantitative researchers study and consider before collecting or analyzing interview data.

On the other hand, there are some interview methods that are considered more often by qualitative than quantitative researchers. These include the way culture, emotional tone, and speech or linguistic differences influence interviewer question delivery, participant question interpretation, and researcher or coder interpretation of participant responses.

Consider that quantitative researchers have increasingly turned to statistical methods, such as latent class analysis, to model heterogeneity. As a result, the quality of quantitative family research has improved. What effect does participant heterogeneity have on interview data? Can we afford to assume that participants interpret questions uniformly or that interviewers deliver them reliably? Have we established that interviewer uniformity improves the reliability and validity of interview data?

It seems likely that we can improve family interview data collection if we study and draw from the full spectrum of disciplines and approaches - from ethnography to multivariate statistics. There are many unresolved interviewing methods issues, particularly for researchers interested in family systems who collect data from multiple family members.

Here is an interesting - and perhaps provocative - passage on this issue taken from Mishler (1991), Research interviewing: Context and narrative, Cambridge, MA, Harvard University Press.

"Lazarsfeld (1935), one of the great pioneers of survey research, understood that variability in how interviewers ask questions is the key to good interviewing and not a problem to be solved by standardization. He recommends a different approach than appears to have been adopted by successive generations of researchers. He refers to the ''principle of division," the aim of which is to adapt "the pattern of our questionnaire to the structural pattern of the experience of the respondent" (p. 4). He recognized that the attempt to fit the questions to respondents' different experiences was, even then, in conflict with usual procedure and traditional opinion for questions to be worded in the same way for all respondents. Instead he argues, "we advocate a rather loose and liberal handling of a questionnaire by an interviewer. It seems to us much more important that the question be fixed in its meaning, than in the wording" (p. 4). Of course, to follow this principle would require tape-recording interviews so that the "meaning" of the questions asked by different interviewers could be determined; as we have seen, such studies are rare."

Posted by vonko002 at 8:37 AM