Implications relevant to CFA model misfit, reliability, and the five-factor model as measured by the NEO-FFI

Gilles E. Gignac, Timothy C. Bates, Kerry L. Jang

Research output: Contribution to journalArticlepeer-review

Abstract

Several attempts to model the NEO PI-R/NEO-FFI as a FFM have failed to achieve adequate CFA model fit, which has prompted some personality researchers to suggest serious problems with SEM. In this paper, it was hypothesized that a portion of the model misfit associated with the NEO-FFI may be due to intra-dimensional correlated residuals, which was expected to result in upwardly biased estimates of internal consistency reliability. Based on a latent variable modeling approach to the estimation of internal consistency reliability, the degree of reliability overestimation was found to range from .02 to.06. Because the consequences can have implications for internal consistency reliability estimation, it is recommended that researchers attend to evidence of model misfit, rather than dismissing the issue as a problem with SEM. (c) 2007 Elsevier Ltd. All rights reserved.

Original languageEnglish
Pages (from-to)1051-1062
Number of pages12
JournalPersonality and Individual Differences
Volume43
Issue number5
DOIs
Publication statusPublished - Oct 2007

Keywords

  • confirmatory factor analysis
  • personality
  • NEO-FFI
  • model-misfit
  • CONFIRMATORY FACTOR-ANALYSIS
  • PERSONALITY-INVENTORY
  • BIG 5
  • COEFFICIENT ALPHA
  • SCALE

Fingerprint

Dive into the research topics of 'Implications relevant to CFA model misfit, reliability, and the five-factor model as measured by the NEO-FFI'. Together they form a unique fingerprint.

Cite this