Page 362 - Contributed Paper Session (CPS) - Volume 7

CPS2130 Abdul-Aziz A. Rahaman et al.
estimation method adopted by the first three residual estimators which
2
yielded different residual parameter estimates. The EM method gave (df
=23, N=145) = 57.80, p < 0.001, RMSEA=0.023, CFI=0.984, SRMR=0.020.
2
Although both and RMSEA indicated worse model fit, the impact on CFI
was small and SRMR actually indicated better model fit compared to the other
three existing methods. Therefore, it implies that with these methods, the
model fit indices gave ambiguous fit information. As a result, the standard
errors associated with the parameters and the comparative fit information
would be preferable in understanding which specific residual estimator
method gave a better parameter estimate. Thus, the fit information using EM
method was comparable with the other three existing methods. Hence the
AIC, BIC, and CAIC all strongly favoured the EM method over these other
methods. Moreover, the parameter estimates in Table 1 above, indicated that
whereas estimates were strongly close for these existing methods, they were
more robust with Bartlett’s and in particular the EM method.

4. Conclusion
In conclusion the strength of the existing methods are the weaknesses of
EM method, and vice versa. It was therefore found from the comparative
model fits information, by comparing among the three existing residual
estimators, that the Bartlett’s based method gave better residual parameter
estimates over the regression-based method and the Anderson Rubin based
method. However, the EM method gave better residual parameter estimates
than the other three existing methods (i.e. the regression, Bartlett’s and the
Anderson Rubin based methods).

349 | I S I W S C 2 0 1 9