Related Experiment Videos
Evaluating the MSCEIT V2.0 via CFA: comment on Mayer et al. (2003)
1School of Psychology, Deakin University, Burwood, Victoria, Australia. gignac@deakin.edu.au
Emotion (Washington, D.C.)
|June 29, 2005
Summary
Reanalysis revealed errors in Mayer et al.
Area of Science:
- Psychological measurement
- Statistical modeling in psychology
Background:
- Critique of confirmatory factor analysis (CFA) in emotional intelligence research.
- Examination of fit indices (normed fit index, Tucker-Lewis Index, root-mean-square error of approximation) accuracy.
Discussion:
- Identified substantial errors in previously reported CFA fit indices.
- Reanalyzed subscale intercorrelation matrix to correct fit index values.
- Evaluated model fit based on corrected indices, differing from Mayer et al.'s conclusions.
Key Insights:
- The 1-factor and 2-factor models demonstrated poor fit upon reanalysis.
- The 4-factor model, while initially appearing to fit well, presented issues.
- A non-positive definite matrix in the unconstrained 4-factor model indicated factor collinearity.
Outlook:
- The collinearity of Perceiving and Facilitating factors suggests a 4-factor model is implausible.
- Highlights the importance of rigorous statistical analysis and reporting in psychological research.
- Underscores the need for careful interpretation of model fit, especially concerning factor collinearity.