3rd December 2009 at 10:17 am #5305Kristian KarlsonMember
I study and use quantitative sociology and do not have much knowledge about qualitative methods. However, I often encounter an argument that researchers using qualitative methods (not all but some) have against random sampling. They say that since they are not interested in generalizing their findings, it is not necessary for them to do random sampling of the group they study.
While I agree with such argument, it still has some limitations, because random sampling has other advantages besides that of the possibility of generalization. In fact one might argue that it is not the property of generalization that is the most important feature of random sampling! I would like to start a thread that discusses other desirable features of random sampling and why it is / is not relevant for methodology in general (whether quantitative, qualitative, or mixed). I have two comments to start the debate:
First, the “nice thing” about random sampling is that you do not need to worry about your results being attributable to the way you selected your informants/respondents. As I see it, this is a fundamental logical problem which is not only specific to quantitative methods, but all methods. Random sampling is one way of avoiding the problem. As I see it, that must be desirable for researchers using qualitative methods as well as other kinds of methods.
Second, simple random sampling has another nice feature, namely that each informant/respondent conveys the same amount of information to the analysis. In more technical terms, they each have weight 1/N in which N is the total number sampled (= the sample size). In other words, I do not have to consider which informants give me the “best” or “deepest” accounts of whatever I am studying. Rather, I can––without hesitation––accept all answers/accounts as having the same significance for my study. Ain’t that wonderful?
What do you think? Isn’t random sampling wonderful, or? Hope to get some responses!
All the best,
- You must be logged in to reply to this topic.