Political Ideology and Issue Importance

Abstract

Past research has shown that issues vary significantly in their salience across citizens, explaining key outcomes in political behavior. Yet it remains unclear how individual-level differences in issue salience affect the measurement of latent constructs in public opinion, namely political ideology. In this paper, we test whether scaling approaches that fail to incorporate individual-level differences in issue salience could understate the predictive power of ideology in public opinion research. To systematically examine this assertion, we employ a series of latent variable models which incorporate both issue importance and issue position. We compare the results of these different and diverse scaling approaches to two survey data sets, investigating the implications of accounting for issue salience in constructing latent measures of ideology. Ultimately, we find that accounting for issue importance adds little information to a more basic approach that uses only issue positions, suggesting ideological signals for measurement models reside most prominently in the issue positions of individuals rather than the importance of those issues to the individual.