Message/Author 

arif özer posted on Monday, January 19, 2015  3:39 pm



I used WLSMV estimator when I performed CFA. I have two model. One of them is twofactor model (chisquare= 43.41, df=26) and the other is onefactor model (chisquare= 44.16, df=27). I applied chisuare difference test by using mplus guide. chisquare difference test = 62.18, df= 1; p=.00. How it can be obtained!.. while there is only one point between two model. Thanks 


Which version of Mplus are you using? 

arif özer posted on Monday, January 19, 2015  4:50 pm



Mplus 6.12 


Please send the 2 outputs to Support along with your license number. 

arif özer posted on Tuesday, January 20, 2015  10:31 am



I bought the program as part of a project three years ago. As soon as I go the school, I'II send license number. 

arif özer posted on Tuesday, January 20, 2015  11:38 am



two output can't be send because of KB limit. Is there another posting way? 


You can zip it. Or, rerun with less output options requested. 

arif özer posted on Tuesday, January 20, 2015  1:36 pm



1) Mplus VERSION 6.12 MUTHEN & MUTHEN 01/20/2015 2:45 AM INPUT INSTRUCTIONS TITLE: CFA with categorical factor indicators DATA: FILE IS kf.dat; VARIABLE: NAMES ARE M1M9; CATEGORICAL ARE M1M9; MODEL: Ozfark BY M1 M2 M3 M4; Kulfark BY M5 M6 M7 M8 M9; SAVEDATA: DIFFTEST IS deriv.dat; OUTPUT: STANDARDIZED; Estimator WLSMV THE MODEL ESTIMATION TERMINATED NORMALLY MODEL FIT INFORMATION Number of Free Parameters 39 ChiSquare Test of Model Fit Value 43.408* Degrees of Freedom 26 PValue 0.0175 RMSEA (Root Mean Square Error Of Approximation) Estimate 0.058 90 Percent C.I. 0.024 0.087 Probability RMSEA <= .05 0.309 CFI/TLI CFI 0.979 ChiSquare Test of Model Fit for the Baseline Model Value 880.214 Degrees of Freedom 36 PValue 0.0000 WRMR (Weighted Root Mean Square Residual) Value 0.626  

arif özer posted on Tuesday, January 20, 2015  1:41 pm



2) TITLE: CFA with categorical factor indicators DATA: FILE IS kf.dat; VARIABLE: NAMES ARE M1M9; CATEGORICAL ARE M1M9; ANALYSIS: DIFFTEST IS deriv.dat; MODEL: Ozfark BY M1 M2 M3 M4; Kulfark BY M5 M6 M7 M8 M9; Ozfark WITH Kulfark @ 0; SUMMARY OF ANALYSIS Number of groups 1 Number of observations 200 Number of dependent variables 9 Number of independent variables 0 Number of continuous latent variables 2 Estimator WLSMV MODEL FIT INFORMATION Number of Free Parameters 38 ChiSquare Test of Model Fit Value 250.591* Degrees of Freedom 27 PValue 0.0000 ChiSquare Test for Difference Testing Value: 62.184 df: 1 P: 0.0000 RMSEA Estimate 0.203 90 Percent C.I. 0.181 0.227 Probability RMSEA <= .05 0.000 CFI 0.735 WRMR 2.223 


Your postings are for two 2factor models which only differ by a factor covariance being zero or not. Your initial question was about a 2factor vs a 1factor. Also, we request posting in only 1 window. Longer issues should go to Support. 

arif özer posted on Tuesday, January 20, 2015  2:03 pm



Ahem!... execuse me ı try to use this forum. And thank you very much for your interest and reply. My mainly interest is to compare 2factor vs 1factor. To solve it: is it adequate that f1 WITH f2 @ 1; 


That's one approach sometimes usedm, if the WITH statement is referring to the correlation. But it is not a fully legitimate chi2 difference test because corr=1 is on the border of the admissible parameter space. 

Back to top 