ABSTRACT
This study uses survey research to develop measures of social change in Taiwan. My 1991 Taipei survey was designed as a repetition of questions asked in Taipei in 1963. It is one of a family of longitudinal designs in which sample survey data are collected on the same set of variables at two periods of time, using nonidentical but comparable cases in each period. The 507 respondents in 1963 and the 545 respondents in the 1991 sample are comparable in that they are all male, Taiwanese, household heads, and living in Taipei city at the time of the interview. The distribution of the two samples on key sociodemographic variables such as education, income, occupation, and household size is significantly different, because they reflect the real changes that have occurred in these variables over the 28-year period. These are treated not as methodological problems of nonidenticality between the samples, but as representative of the real substantive changes that have occurred, and are analyzed in detail in chapter 3. At the individual level, we can compare the behavior or attitudes of Taiwanese who in both the 1963 and the 1991 samples are, say, under 40 years old, or have a given level of education, or occupation. But the aggregate behavior or attitudes being studied may well differ between the two samples as a whole because the proportion of respondents in the given age group, educational, or occupational category has changed significantly. Since this is precisely what this research is about, there would be no point in designing the 1963 and 1991 samples so that they would have the same distribution on each of these sociodemographic variables.
