Classical test theory based and factor analytic structural models for analyzing change: A note on differences and similarities |
| |
Authors: | TENKO RAYKOV |
| |
Affiliation: | Department of Psychology, University of Sidney, Sidney, Australia |
| |
Abstract: | This note is concerned with differences and similarities between structural models for analyzing change, which are conceptualized within two different modelling traditions: the one based on the classical test theory, and that within the factor-analytic approach. It is shown that these two possibilities lead to models for studying change, which are indistinguishable when using for data analytic purposes structural modeling programs, such as LISREL, EQS, COSAN, LISCOMP, RAMONA, EzPATH, SAS PROC CALIS. The reason for this data-analytic equivalence of the two conceptually different types of models is the confounding of their differences in the corresponding implied covariance matrix structures. |
| |
Keywords: | Measuring change classical test theory factor analysis structural equation modeling confounding |
|
|