Measuring Discrepancies Using Latent Variable Modeling
Lisa M. Yarnell, University of Southern California
Toni Falbo, University of Texas at Austin
This paper aims to describe an innovative use of a new approach to latent variable modeling that involves latent difference scores (LDS: McArdle, 2009), reflecting the degree of difference between subjective self-views and more objective assessments of the self. We created latent difference scores reflecting the degree of accuracy across the domains of intelligence, general health, body mass, and attractiveness, using data from the third wave of the National Longitudinal Study of Adolescent Health (Add Health). We tested structural models examining the consequences of self-perception accuracy on a variety of socioeconomic and mental health outcomes assessed roughly six years later. What is innovative about our use of the LDS framework is that we use it to model intra-individual discrepancies at the same point of time. This approach avoids the weaknesses of prior methods because measurement error is removed from the latent difference score.
Presented in Poster Session 7