Academic journal article Ethnology

Testing Theory and Why the "Units of Analysis" Problem Is Not a Problem

Academic journal article Ethnology

Testing Theory and Why the "Units of Analysis" Problem Is Not a Problem

Article excerpt

This article discusses why cross-cultural comparison is possible and why theory needs to be tested universally. It discusses why worldwide cross-cultural results are likely to be more generally valid (and more useful practically) than more limited comparisons, and certainly more generalizable or trustworthy than single-case analyses or theory that has never been tested. Answers are provided for the main objections to worldwide cross-cultural research: the supposed incomparability of cultural traits, the supposed incomparability of units of analysis, the supposed impossibility of unbiased sampling, and what is known as "Galton's problem." (Theory, testing, cross-cultural, units)

Compared with most social sciences, anthropology as a discipline has a larger frame of reference. In the language of statistical inference, it has a bigger sampling frame. The whole world of humans, past and present, is its frame of reference; anthropologists are interested in every place and time that has ever had a human population. The other social sciences generally have more parochial sampling frames. They also generalize about human behavior, but their data are drawn from more restricted domains, certainly not the whole of human prehistory and history. So, for example, researchers in some disciplines (sociology, economics, political science, psychology) are mostly concerned with one or another complex commercial society, typically their own. Or they may deal just with the nation-states of the past few hundred years (political science, history, international relations). Even those psychologists who call themselves "cross-cultural psychologists" (who are aware that humans are not exactly the same psychologically in all societies) generally do their research in countries that have schools and colleges because the tests and measures they want to employ require subjects who can read and write. Years ago, Otto Klineberg (1980:v) scolded his psychological colleagues for their ethnocentrism: "How could psychologists speak of human attributes and human behavior when they knew only one kind of human being?" We can say the same thing to most social scientists (other than anthropologists). Are literate subjects representative of all human beings? Is the nation-state the only kind of human polity? Do commercial and industrial economies represent all the possible ways to make a living?

This is not to say that anthropologists are never parochial. Some anthropologists generalize from one society to a set of societies or even to all of humanity. This kind of parochial generalization can be called "other-centric," in contrast to the kind of parochialism we call "ethnocentric," which glorifies one's own society. Then there is extreme relativism, which is not even parochialism: it rules out the possibility of any generalizations. For an extreme relativist, understanding can only come from analysis of a single case; comparison is unjustifiable, even hegemonic or imperialist. But if comparison is possible and even desirable (which is what we try to show here), extreme relativism should be interpreted as a variant of what philosophers call solipsism, the belief that the self can be aware of nothing but its own experiences and states. Described so baldly, extreme relativism looks silly.

There is an irony here. Anthropology as a discipline claims the whole world as its domain; all of humanity past and present, in all times and places, is of interest and concern to it. But there are many anthropologists who are resistant to comparisons of any kind, out of extreme relativism or just because they are unsophisticated about how science operates. If generalizations are suggested, they are often the kinds of parochial over-generalization (other-centric, ethnocentric) mentioned above. In the absence of systematic supporting evidence, why should anyone believe that any one human population is like all others, or even like all others of its type? We might like to think that the ! …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.