Cultural classifications under discussion latent class analysis of highbrow and lowbrow reading

Abstract

Cultural classifications, in the sense of categorisations of cultural goods and practices, are often abused in cultural stratification studies to classify people involved with these goods and practices. In this paper the issue is discussed of how to use cultural classifications without begging the question of their definition and without claiming to have at our disposal an archimedean point permitting the classification of cultural goods. Using time budget data from the 1990 Dutch Time Budget Survey (over 3000 respondents) and focussing on one cultural activity (reading in leisure time) latent class analysis - a method for finding the latent classes or categories of variables - provides an empirical basis which permits exploring an extrapolation of Peterson and Simkus's (1992) suggestion (on omnivorous and univorous musical preferences) with respect to actual reading behavior. It is shown that present-day Dutch society is composed of clearly distinguishable categories: non-readers, high readers, and low readers. Besides, there is a small group which combines apparently heterogeneous categories. (Focus on a single cultural practice must make one hesitant to label this group ‘reading omnivores’.) Subsequently, a multinomial logit model is described and used to test hypothesized causal relationships between kinds of reading behavior and a number of background (SES) and intervening variables. These variables help to account for class membership and thus make possible the explanation of differences in reading behavior.