William R. Catton

Changing Cognitive Structure as a Basis for the “Sleeper Effect”

Let me summarize this general discussion of statistical inference briefly. Area data alone were considered. Three kinds of data resulting from data collection processes were identified. Two kinds of theoretical interpretations of these data were posed, and the remainder of the paper stressed the interpretation in term of aggregates as specified by a particular definition of structure and a theoretical statement of social structure. Four major problems of statistical inference were identified. Various approaches to these inference problems were examined critically. In some cases specific statistical procedures were recommended for use, while in other cases certain difficulties in present procedure were indicated but solutions were not presented. It is the intention of this paper and the research upon which it is based to aid in the development of sound statistical procedures and appropriate statistical inference in the analysis of area data. Illustrative computations of all these procedures are available in the authors unpublished doctoral dissertation. Manuscripts are in preparation which will bring these computations to the attention of a wider audience and thereby encourage the use of more appropriate statistical methods in studies analyzing area data.