Metablog

On Framing Questions

One of the dangers of dealing analytically and academically with topics with which we're personally familiar is the inclination to reason from intuition or personal experience. Topics like privacy, for example, tend to yield a lot of reactions of the form "oh, but what I prefer is..." Similarly, as readers and writers of blogs, we are tempted to take our own experiences as a primary source of information, and then see other accounts relative to that.

There are many dangers in this. One is that our own experiences are, by definition, a degenerate sample (even if we think that we're quite "normal") -- especially true when computer scientists talk about phenomena of computer use. The second is that we are often not truly conscious of our own practices (the reason that we normally observe practice rather than just taking an informant's word about what they do.) The third is that, pretty much by definition, we generally lack perspective on our own practices -- the same reason that anthropologists don't rely on a single perspective, and why some concerned with qualitative methodology, such as George Marcus, have vociferously argued for "multi-site ethnographies."

Now, that's not to say that personal experience isn't valuable, and that intuition should be ignored. The issue is not to allow it to play any primary role. In fact, in any kind of qualitative work, we must work with our own intuitions, either to bracket particular interpretations of data, or to help identify questions that should be investigated further.

The question is how to use intuition to frame scientific questions. Our own experiences don't answer questions, but they can help us find out what questions to ask. When personal experience differs from that described in a study, we can use this different to frame empirical questions -- how widespread are these differences, what is the range of different practices, what different communities might encounter a technology, etc -- or to frame analytic questions -- what underlying factors might account for different uses or experiences, how are different approaches structurally related to each other, what further experiences might be expected, etc. Similarly, when our experiences tally, we should nonetheless use this to question the data -- by asking, for example, whether the similarity between our experiences and reported experiences might lie in certain common background factors which may turn out to be different in other circumstances.

I think it's important that we try to bear this stuff in mind when we are exploring questions of computer-mediated communication, blogging technologies and behaviors, styles of Internet use, personal disclosure, privacy norms, journalling practices, etc. One thing that I will insist on more in future sessions than I did this week is to maintain a scientific/experimental orientation towards the material, and not simply to juxtapose personal with reported experience (or, worse yet, to grant one primacy), but to use it to frame questions for further investigation (perhaps in the papers/projects later in the quarter.)