Ethnography as a Research Method
Orr stated that there are “three characteristics of ethnographic description”: interpretation, interpretation from social discourse and preservation of the discourse. Ethnographic research is then the observation of how socially reality is constructed. A benefit to this approach is thick description, as evidenced by Orr’s recounting of conversations and the intimacy with which the team members share with the research. Often, Orr’s subjects are conversing with him and even when they are not, the act of being observed is a new factor injected into the reality of the way the technicians and customer act. As a result, observations are biased. How does this bias affect the outcome of what is observed? Does observation without the awareness of those being observed mean a less biased observation? How does this bias compare to the bias of running regressions to observe correlations, for example? Is one better than the other?
posted by Anonymous at 4:53 PM
3 Comments:
About observer subjectivity:
>>>How does this bias affect the outcome of what is observed?
We could go about this question in several ways. We could talk about what kind of a bias he had, and how certain kinds of bias can result in different kinds of findings. Imagine, for instance, if Orr had been a high-schooler or an IRS agent. But on another level, isn’t bias inevitable? Put differently, is it ever possible to do ethnography “objectively”? I think that if we take Searle’s thesis of socially constructed reality, we have to submit ourselves to the inherent subjectivity in research. But then again, to resurrect the questions we had with Weber, what is then the purpose of social science research if the Truth is not be had.
>>>Does observation without the awareness of those being observed mean a less biased observation?
If we frame this question within Searle’s thesis, then there is no observation without observer. That there is observation already instantiates an observing entity. So, say observation is done in secret and the “subjects” studied go about more “naturally”—wouldn’t the reported observation still be through the (subjective) lens of the observer?
>>> Implicit in this is that he wasn’t letting his own experiences interfere with his project because his work was sufficiently different. Did he succeed? [by KSG]
Maybe one way to respond to this question is through the notion of “bricolage.” A bricoleur is someone who works with whatever she finds available at hand. Orr ended up in this project more or less by chance, and, further, in an important sense he didn’t *pick* to become an ethnographer in Xerox world. Orr sort of just “found” himself in this work. In this sense Orr was a bricoleur (so were the technicians who had to deal with whatever problem du jour). But a part of being a bricoleur is not having directive control on the cause, means and goals of your work. Given this sort of flailing embeddedness within the context, the possibility of mitigating his influence becomes even less likely if not entirely impossible.
Let me toss in a distinction that might be useful here. "Description" might mean one of two things:
a) pictorial representation of an externally-existing reality; or
b) an account from the margins, translating between inside and outside.
The former is the kind of thing for which descriptive statistics are designed, and issues like representativeness, sample size, and "accuracy" are appropriate. Think here of a public opinion poll -- we can't know whether Bush's approval rating in the general population is 25% or 75% unless we follow certain procedures in randomly selecting informants and then aggregating the data. There's no causal analysis here, just description.
But the latter, which we could follow Orr in defining as ethnographic description, is a very different animal. I don't see in Orr (or in other ethnographers) much of a claim to "accuracy" in the same sense that statistical descriptions claim. Instead, I see something like what we find in literary criticism: "here's what I made of this text or these texts," rather than "here is what this text or these texts mean." Hence the evaluationa criteria for such a description or interpretation might be different: is the interpretation defensible, rather than accurate.
Now, even if one buys this distinction between two types of description, this doesn't necessarily get us very far in terms of debating the value of each of the types. And that question of course intersects with the distinction itself, since if one only sees value in one of the two types one is likely (I think) to question the distinction in the first place, and argue that all description is really one or the other type.
Just stirring the soup from abroad.
Post a Comment
<< Home