Profgrrrrl who (1) I should read more regularly and (2) I assume to be a sociologist based on this statement:
[I do] ... mixed methods work, in a field where all too frighteningly often I hear people describe themselves as either qualitative or quantitative researchers
Quite so. It's clear to me from reading books like Howard Becker's Tricks of the Trade, and King, Verba and Keohane's Designing Scientific Inquiry that the problem of research design for accurate inference is common to all areas of the social sciences.
Social scientists essentially ask the question: "Why do people act the way they do?" There are variations on how you specify that question, but it's the basic question at the heart of all social science research. [he says humbly] How do people form expectations about how the price of goods and services change [inflation]? What makes people sexually attracted to others? Why did the assassination of a relatively minor Austrian archduke lead to a bloody and costly 4 year war that dragged in countries not even remotely near Sarajevo? You can assign disciplines to all these questions, if you like, but the form of the question is the same.
The data we have to answer these questions is often incomplete, either because (1) it would be prohibitively expensive to ask everyone what they thought about the issue, or (2) things get lost. The latter is the historians problem.
So we make the leap from incomplete data to attempting generalizations about the behavior in question. The formal structure of this inference is much clearer for quantitative research, though that is not to say that everyone always remembers what they're doing and the problems inherent in the exercise.
Moreover, data are not inherently qualitative or quantitative. The researcher makes an active choice about how to analyze them. Some data go better with some analyses than others. And new insights sometimes come from approaching the same old data with a different analysis.
To take an example from my own research, I tend to think about the dissertation as involving three primarily quantitative chapters, one primarilyqualitative chapter, and one (the introduction) which is a bit of a mixture. But that's a misnomer. In chapter two I ask "how did the income earned by husbands affect whether wives worked, and how did that change from 1890-1940?" And the numbers give me a certain sense for what's going on -- husband's income had a smaller effect over time. But they mean little-to-nothing without the extra information that's provided by people writing about what they thought about what was happening.
Academics are quite uninclined to study themselves, but I think the persistence of this qualitative/quantitative debate in certain disciplines (sociology especially, history in some areas) is well worth studying in itself ... if it hasn't been done already.
Posted by robe0419 at February 26, 2005 06:10 PM | TrackBackHistorians had something like the navel-gazing debate you call for during the 1970s and 1980s controversies about "cliometric" history. But since the new cultural history defenders of qualitative history seemed to win out (at least institutionally) in those debates, the question doesn't seem to come up as much anymore. Frankly, if historians were still required by professional canons to crunch numbers, I would be in deep trouble.
Posted by: Caleb at February 28, 2005 08:46 AM