
Message/Author 

tino posted on Friday, August 25, 2017  12:47 am



Hello, in the first part of example 12.7, the parameter estimates seem to be predetermined by the researcher using constraints. For example, the cov between f1 and f2 is set to .25. But what is the rationale for using such constraints here  I thought the idea would be to inform the MC simulation by the actual empirical results instead of some predetermined parameter estimates? Wouldn't it also make sense to proceed like this: In Mplus, (1) estimate the model of interest using the real data but without any constraints. (2) Save all estimates from (1) in a *.dat file (3) Use this *.dat file for the subsequent Monte Carlo study 


The values given are the population parameter values to be used in the generation of the data. They could be obtained by your numbers 1, 2, and 3. In the first step of the example, no model is estimated. Data are generated and saved for use in the second step of the example. See Example 12.1 where data are generated and analyzed in one step. 

Back to top 

