Cultural Units in Cross-Cultural Research

In 1889, Galton argued that often traits diffuse across cultures and therefore we could never really know if cultural traits arose independently as adaptive responses or were a result of diffusion. With numerous hypothetical and actual examples we show that, statistically, there are no mechanical fixes to this problem. We conclude by asserting that each cross-cultural research question should find its own solution to Galton's problem and that, in fact, this is not a problem at all, but rather an asset which can be used to trace historical networks. (Galton's problem, cultural units, community, language, megacultural regions, Standard Cross-Cultural Sample) The notion of the cultural unit has two meanings. One meaning considers cultural units as elements out of which culture is composed. There are doubts whether such entities exist at all (see Gatewood, this volume). In any case, this meaning has no direct relevance on cross-cultural research. One can discuss and compare molecular units, such as postmarital residence practices, without being concerned about the elemental units that comprise this practice. The other meaning of cultural units, however, does have direct relevance and must first be descriptively rather than formally explained. To start with, within cross-cultural research the problem of cultural units is not quite identical to the problem of units of comparison (though both problems are connected). The solution to the problem of comparison was solved long ago by Naroll (e.g., 1970), who suggested that the unit of comparison should be community, not culture. However, the problem of cultural units arises immediately upon selecting communities for cross-cultural comparison. As the very notion of cross-cultural research implies, the communities that are used for comparison must belong to different cultures. Clearly, the inclusion in a sample of a number of communities that belong to the same culture could result in the production of a correlation confirming a false hypothesis (or, alternatively, rejecting right ones). This can be illustrated with a fictional example. Our hypothesis is that the consumption of dates (the fruit) enhances the sexual drive among religious specialists, and the consumption of red wine inhibits that drive in religious specialists. Sexual drive is measured by asking these specialists their frequency of sexual intercourse after consuming dates or wine. The frequency of sexual intercourse should go up after consuming dates and go down after consuming wine. To test the hypothesis, imagine a sample of communities described in Table 1: Table 1: Communities in Sample 4 Spanish 4 Iraqi 1 Russian 1 Estonian 1 Javanese 1 Ganda 1 Greek 1 Maronite-Lebanese The statistical analysis of the data for this sample is most likely to produce the following results: Table 2: Dates vs. Red Wine Consumption/ Sexual Intercourse Cross-Tabulation (version 1) Sexual Intercourse Near Absent Regular with Regular with More One Partner than One Panner Wine 4 1 0 No wine/dates 1 1 2 Dates 1 0 4 Total 6 2 6 Value Approx. Sig. Gamma .79 .001 Spearman Correlation .64 .014 Thus the test is likely to support a patently wrong hypothesis. One of the main reasons for this is that the sample included eight communities from two national cultures. One of these (Spain) is characterized by Catholicism (and hence a celibate clergy) and the wide use of red wine. The other is Islamic (which permits polygyny for religious specialists) and has a wide consumption of dates. The solution normally proposed to avoid such problems is simply to include no more than one community using the same language in the sample. …