Abstract
Confirmatory factory analysis (CFA) has been widely used in psychological and educational research. By summarizing and referring to results of three simulation studies, the present paper discussed strategies particularly relevant to small sample studies. Specifically, it was suggested that more items should be used and factor loadings can be restrained to be equal within the same factor. Generally, larger sample is more desirable, but when resource is limited, the above two strategies may solve non-convergence problems and increase the estimation accuracy of important parameters. Furthermore, the use of two indicators within a factor, the use of less than four indicators when sample size is only about 100, and the combination of items into parcels and using their averages as indicators are not recommended because they may lead to non-convergence, large standard error and biased parameter estimation.
| Keywords: | structural equation modelling; confirmatory factor analysis; multivariate analysis; small sample |
|---|
[Chinese Version | Index | Acta Psychologica Sinica | Other Journals | Subscription form | Enquiry ]
Mail any comments and suggestions to
hkier-journal@cuhk.edu.hk .