The Stability and Reliability of Political Efficacy: Using Path Analysis to Test Alternative Models

Abstract

The reliability and stability of survey items designed to measure political attitudes are important to the study of political behavior. Several past studies have examined the reliability and stability of items measuring one construct, that of political efficacy. The results of this prior research have been contradictory, in part because of the limitations of the methodologies used. In this article, the authors employ path analysis to examine more closely the stability and reliability of the four SRC items commonly used to measure political efficacy. The American Panel Study (1956–1960), in which efficacy is measured at two points in time, is used as the data base. The authors conclude that two of the four items (NO CARE and NO SAY) seem to measure best what is meant by political efficacy. These two items are more stable and reliable than previously thought, while the other items are relatively unstable and unreliable, and they display systematic differences from each other and from the NO CARE and NO SAY items.