Subjective Class Identification: A Multiple Indicator Approach

Research on subjective identification has tended to use the same single-item measure of this construct without regard to its validity and reliability. To address these and related measurement issues, multiple indicators of subjective are analyzed within the framework of a path analytic model with unobservable variables. Measures of objective class and political alienation are also included in the construction of an overidentified model, and Joreskog's procedure for the analysis of covariance structures is used to find an acceptable fit to the data. Final models for black and white subsamples support the assumption of unidimensionality which underlies previous measures of subjective identification. Validity and reliability estimates are given for various indicators, including the conventional measure. Also, advantages of the multiple indicators approach are demonstrated by comparing regressions of the standard, single-item measure and a composite measure of subjective on education, occupational status, and income.