Specifically we compared the fit of two models one in

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: rted in other studies (John & Srivastava, 1999). To formally test the structural stability of the Big Five, we used structural equation modeling to conduct an asymptotic test of the equivalence of the two correlation matrices. Specifically, we compared the fit of two models, one in which the intercorrelations among the Big Five were freely estimated at each time point and the other in which the intercorrelations among the Big Five were constrained to be equivalent across the two assessments. If the unconstrained model fit significantly better, this would indicate significant change in the structural relations among the Big Five. To conduct these analyses, we analyzed the matrix of covariances among the 10 scale scores. We specified a single-indicator latent variable model with one latent variable associated with each scale score. We identified the model by fixing the variances of the latent variables to 1 and the variances of the residuals to 0. We then estimated paths among all latent variables and between each latent variable and its indicator. In this analysis, the matrix of covariances among the latent variables is equivalent to the matrix of correlations. This leads to a fully saturated model (CFI = 1.00). We reestimated the model after placing 10 pairwise equality constraints between paths at Week 1 and Year 4. A chi-squared difference test indicated that constraining the model did not lead to a significant reduction in fit, χ2∆ (df = 10) = 8.5, ns CFI = .99. Thus, the intercorrelations among the Big Five were highly stable over time. 630 Robins et al. Ipsative Stability of Personality To explore the patterning of traits at the individual level, we compared the stability of individual trait configurations over time using Cronbach and Gleser’s D2, D′2, and D″2 indices. The distribution of D2 values ranged from .04 to 8.24, with a mean of 1.63 and a standard deviation of 1.24 (the 25th, 50th, and 75th quartile scores were .77, 1.31, and 2.21, respective...
View Full Document

This note was uploaded on 12/02/2010 for the course PSYCH PSY BEH P2 taught by Professor Susanturkcharles during the Fall '10 term at UC Irvine.

Ask a homework question - tutors are online