View article

210 PUTTING IT ALL TOGETHER such as “disagree,”“somewhat disagree,”“somewhat agree,” or “agree.” The resulting data emerge already coded into usable form for data analysis, and tend to be prized for hypothesis testing because the results are so easy to compare. The problem is, of course, what if the researcher asked the wrong questions or did not provide an important category of answer? The results may be reliable across in for mants without being valid. Although they are more easily replicable than other types of interviews, survey interviews with close-ended responses may miss the point and may overlook important opportunities for learning from the research subjects.

On the other side of the interviewing world, researchers in fields such as anthropology and sociology have written extensively about ethnographic or interpretive methods of interviewing (eg, McCracken 1988; Spradley 1979; Werner and …