jeon small posted on Friday, March 01, 2013 - 8:05 pm
How do you perform a posthoc power analysis. I have a sample of 438 girls and boys. We used 11 indicator variable to form four variables. The fit of the structural model was satisfactory with CFI=.954, RMSEA=.05, DF=40 CHI SQUARED=95.85 PROBABILITY=.0000 The standardized regression weights were significant and ranged in size form .65-.03.
See Example 12.7 where the values from a real data analysis are saved and used as population parameter values for a Monte Carlo simulation study that can examine power. This method has its critics. You may want to explore this issue on SEMNET.
jeon small posted on Monday, March 04, 2013 - 1:38 pm
I ran this model in Amos using maximum-likelihood estimation (for the missing data)--however, I can not get the model to converge in Mplus. What am I doing wrong? My output is below.
STRUCTURAL MODEL OF ADOLESCENT ALCOHOL USE AND PERCEPTION OF IPV
SUMMARY OF ANALYSIS
Number of groups 1 Number of observations 553
Number of dependent variables 10 Number of independent variables 0 Number of continuous latent variables 4
Estimator ML Information matrix OBSERVED Maximum number of iterations 1000 Convergence criterion 0.500D-04 Maximum number of steepest descent iterations 20 Maximum number of iterations for H1 2000 Convergence criterion for H1 0.100D-03
Input data file(s) h:/M_russell/mplus10_08_12v1.csv