ABSTRACT
Especially in cross-national comparative research, it is essential that observed variations in survey responses are properly assessed as to whether they capture genuine country differences or whether they are merely manifestations of various methodological artifacts. This chapter applies multiple correspondence analysis and categorical principal components analysis as data-screening devices to a battery of items contained in the 1996 and 2004 ISSP surveys that are collectively referred to as political efficacy and trust variables. Four countries representing different continents, cultures and languages (Sweden, Japan, the United States and the Philippines) were selected for the data-screening analyses. The findings clearly show that the structures of both substantive and methodologically induced variation differ enormously between these four countries. The first implication is that levels of political efficacy and trust cannot be directly compared across these countries. A second implication is that, even for data of high quality such as the ISSP surveys, it is imperative that multivariate data-screening procedures be conducted in order to determine the conditions under which countries can be compared.
