To illustrate the multiple-indicator CTCU and CTC(M-l) models, we applied them to the same data set as the single-indicator models and included the second test halves of the scales as second indicators (see Table 20.3). In the multiple-indicator models, the first indicators were identical to the ones analyzed in the single-indicator models. This makes it possible to show the differences in the estimated parameters, particularly the reliabilities between the single-indicator and the multiple-indicator models. In the applications we assumed that the two indicators are homogeneous indicators of each trait-method unit. Thus, the loading parameters of the two indicators were set equal to each other. The fit of the two models was very good (CTCU: x2= 127.32, df= 123, p = .38, CFI = 1.00, RMSEA = .01; CTC(M-l): %2= 123.08, df = 120, p = .41, CFI = 1.00, RMSEA = .01). The fit of these models is in the same range as the fit of the MTMM

Was this article helpful?

0 0

Post a comment