1983
DOI: 10.1207/s15327906mbr1802_2
|View full text |Cite
|
Sign up to set email alerts
|

Cross-Validation Of Covariance Structures

Abstract: This paper examines methods for comparing the suitability of alternative models for covariance matrices. A cross-validation procedure is suggested and its properties are examined. To motivate the discussion, a series of examples is presented using longitudinal data.

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1
1

Citation Types

14
506
0
5

Year Published

1998
1998
2014
2014

Publication Types

Select...
9
1

Relationship

0
10

Authors

Journals

citations
Cited by 742 publications
(525 citation statements)
references
References 12 publications
14
506
0
5
Order By: Relevance
“…Model fit was assessed with the model chi-square, the standardized root mean squared residual (SRMR; (Bentler, 1995;Joreskog & Sorbom, 1981), the root mean squared error of approximation (RMSEA; (Browne & Cudek, 1993;Cudek & Browne, 1983), the comparative fit index (CFI; (Bentler, 1990), and the non-normed fit index (NNFI; (Bentler & Bonnett, 1980;Tucker & Lewis, 1973)). The chi-square test is known to be highly sensitive to sample size, and can result in rejection of closely fitting models in large samples.…”
Section: Resultsmentioning
confidence: 99%
“…Model fit was assessed with the model chi-square, the standardized root mean squared residual (SRMR; (Bentler, 1995;Joreskog & Sorbom, 1981), the root mean squared error of approximation (RMSEA; (Browne & Cudek, 1993;Cudek & Browne, 1983), the comparative fit index (CFI; (Bentler, 1990), and the non-normed fit index (NNFI; (Bentler & Bonnett, 1980;Tucker & Lewis, 1973)). The chi-square test is known to be highly sensitive to sample size, and can result in rejection of closely fitting models in large samples.…”
Section: Resultsmentioning
confidence: 99%
“…Noninvariance is claimed if the χ 2 difference is statistically significant (Byrne 2010). However, the χ 2 difference test represents an extremely stringent test of invariance, given that SEM models are at best only approximations of reality (Cudeck and Browne 1983;MacCallum, Roznowski, and Necowitz 1992); thus, we decided that it would be more reasonable to base invariance decisions on a difference in CFI values exhibiting a probability < 0.01 rather than to base such decisions on Δχ 2 (Cheung and Rensvold 2002). Because there is still no consensus on which tests of invariance better represent the phenomena (Byrne 2010), we report both the χ 2 difference and CFI difference results when reviewing the results pertinent to cross-validation in this article.…”
Section: Tests For the Invariance Of A Causal Structurementioning
confidence: 99%
“…78 For this third stage of the pilot study the revised instrument was administered to 116 graduate and undergraduate students (yielding 114 usable questionnaires) in a Southern metropolitan university (Table 1, P6). This was sufficient for a small measurement model to be tested using structural equations modelling (SEM).…”
Section: Sexmentioning
confidence: 99%