[quoting Buckner, 25 Mar 97] After thinking this over, I can't see a problem with the research question Walt refers to. My problem with the scenario is the description of the types of data collected/reported. Personally, I would love to read a case study that utilized an emergent design by a researcher whose goal was true engagement of the learners rather than polite responses to what the learners thought the researcher might want to hear. In my opinion and for this topic in which there is so little quality research, the questions should come from the learners/participants rather than from the researcher.
As far as I'm concerned, Terri has hit the nail on the head with this paragraph. I'd even take it one step further: Rather than bothering with interviewing the participants (learners) the researcher more often than not produces a multiple choice questionnaire which yields the responses (s)he is looking for. As Terri intimates, this means the learner's real thoughts and feelings on the subject are irrelevant to the researcher and her/his theories.
The researcher must be actively involved with the learners (and so also become a participant--another can of worms) to test whether the theory has any relevance.