Identyfikacja błędów pomiaru w modelach zależności strukturalnych

Abstract

One of the key issues in methodology of sociological research concerns identification of measurement errors. This analysis starts from historical overview of methods of analysis of this issue. Then some attempt is made to point out major merits and drawback of various approaches to measurement of errors which ens up conclusion that relatively most fruitfull is that based on confirmatory factor analyses and structural models. Using confirmatory factor analysis I try to prove utility of this approach showing that it allows to increase validity and reliability of measurement. Data come from national samples in Poland and analyses concern effect of social position and attitudes which is one of the most thoroughly explored issues in empirical research,