In trying to fit a polytomous (3-category) IRT model using Mplus (5.1) I have run into the following. To recover the IRT parameters I want to write a script to parse Mplus output. Using the syntax "SAVEDATA: ESTIMATES ARE "estimates.dat";" I am able to write the correct estimates to a file. What I do not get is how the numbers in this file map onto the numbers in the Mplus 'Model results' output.
I assume elements 5-10 of the output are the loadings and that the last 10 elements are the thresholds. However, I cannot deduce what the other values are (i.e. the first 5 elements of 0 and the 0s and 1s in elements 11-28) and could not find anything online or in the manual. Any help would be greatly appreciated.
The file created using the ESTIMATES option is meant for use in a Monte Carlo study. You should use the RESULTS option. It will be shown in the output what is saved and the format. The parameter estimates and standard errors are saved according to the order of the parameters in TECH1.
I am wanting to save the (26x26) tetrachoric covariance matrix from EFA. I am using the following commands:
OUTPUT: SAMPSTAT; SAVEDATA: SAMPLE IS covrem.txt;
While I can get the lower-triangular covariance matrix by this process, it is saved in a manner that does not look like the output from Mplus. (It is saved as 5 columns, 70 or so rows, plus one element on the last row.)
Is there a way to change the format for how this matrix is saved? Ideally, I would like it in matrix format (26 by 26) in a txt document, so that I can then import this matrix into another statistical program. But, I'm open to other options that would make the matrix easier to import into other programs as well.
Is there any way to create an ASCII file that looks more like the Mplus output file? More specifically, I'm looking for a way to create an ASCII file that contains: 1. the point estimates in one column, 2. the standard errors in one column, 3. the test stat in one column, 4. pvals in one column, and 5. a description of the estimates (e.g., F1 on Item1) in next.
I suspect the answer to this question is no (based on my review of the SAVEDATA command documentation), but wanted to confirm that is the case.