ANALYZING PAIRS IN CROSS‐NATIONAL SURVEY RESEARCH

Publisher Website
Google Scholar
Abstract
One device for studying political generations and political socialization is the use of parent‐offspring sampling designs. A key advantage of such linkage designs is that they permit two types of analysis. First, aggregates may be compared, thus allowing an assessment of generational or group correspondence and congruence. Second, the pairs can be used as the unit of analysis for assessing the degree of pair or lineage correspondence. Levels of pair and group correspondence are not necessarily identical for the same set of data. Examples of four major combinations of correspondence levels were found in a study of Austria, Great Britain, the Netherlands, West Germany, and the United States.Among the problems attending pair analysis are the usage of percentage agreement versus correlational techniques and the tendency to view pair correspondence as static rather than dynamic. Illustrations of these problems come from an American socialization panel study. Problems in studying group correspondence include efforts to specify life cycle, generational, and period effects. Alternatives to simple aggregate comparisons include a variety of regression and analysis of variance techniques as well as simpler devices for depicting generation and lineage gaps. For both pair and aggregate analysis the availability of cross‐national data is an opportunity for theoretical and substantive advancement.

This publication has 13 references indexed in Scilit: