For this week’s exercise, I asked my classmates about their usage practice of Wikipedia.My initial idea is to examine the users’ criteria for identifying high-quality answers in social Q&A sites,which is closer to my major area. However, it turned out to be the first interviewee I picked up never use social Q&A and thus didn’t understand what I was asking. This reminded me that delivering your interview question to the right population is the first thing you need to be sure, especially if you would like to ask questions about attitudes, beliefs, behaviors, and feelings.Everyone may have experience to go grocery shopping but obviously just a small group of people uses social Q&A!
Once I changed my topic to Wikipedia, the “population” problem still hasn’t been solved completely due to the high level of homogeneity in subjects. I got quite similar reflections from all three interviewees: they all used Wikipedia as a “starting point” and used the references and external links as indicators to evaluate the quality of certain article (the validity dimension), and mainly focused on the “objective” information the article provided. Dawn said “it will be an issue if any phd student trusts the Wikipedia article too much and does not care what the resources it referenced!” That’s the reason why all the interviewees reached high degree of agreement–because all of them are phd students. However, such bias in the population should be avoided in a research project. I didn’t have chance to share my experience in class about the interview I’m currently conducting –to examine the multilingual searching behavior of Chinese, Shu and I are trying our best to select our interviewees based on the communities they represent (e.g. Chinese lives in different parts of the world) and their professions. The purpose of the selection is to avoid the bias in the population I mentioned above.
As I mentioned in class, I feel unstructured interview turned out to be the most effective one. I assumed very few people will actually look at Wikipedia’s history or discussion pages so I didn’t include related questions in my semi-structured interview with Cheryl. However, in the unstructured interview with Tim he mentioned in some cases he will check those pages and use them as indicators of the quality. This provided new perspectives from the interviewee’s side and helped me to add questions to my interview instrument.