Document Type
Research Paper
Abstract
Confirmatory factor analysis is often used for combining the reports of multiple informants, demonstrating convergent and discriminant validity, and estimating factor correlations (cf. Cole, Truglio, & Peeke, 1997). Although chi-square tests of model fit are biased toward rejecting adequate models when observed score distributions are nonnormal (Curran, West, & Finch, 1996), parameter estimates tend to be robust to violations of multivariate normality (Satorra, 1990). Relatively little attention has been paid to robustness of standard errors for testing statistical significance of parameter estimates even though standard errors are central to determining discriminant validity. The purpose of the current investigation was to examine the robustness of various standard error estimation procedures.
Publication Date
1999
Journal Title
Conference Proceedings: Undergraduate Social Science Research Conference
Volume
3
Issue
1
First Page
175
Last Page
183
Copyright
©1999 by the University of Northern Iowa
Language
en
File Format
application/pdf
Publisher
University of Northern Iowa
City
Cedar Falls, IA
Recommended Citation
Marchik, Sheri and Bengston, Hanna
(1999)
"Robust Standard Errors in Confirmatory Factor Analysis: Implications for Child Psychopathology Research,"
Conference Proceedings: Undergraduate Social Science Research Conference: Vol. 3:
Iss.
1, Article 30.
Available at:
https://scholarworks.uni.edu/csbsproceedings/vol3/iss1/30
