search

michaelhallquist / MplusAutomation

The MplusAutomation package leverages the flexibility of the R language to automate latent variable model estimation and interpretation using Mplus, a powerful latent variable modeling program developed by Muthen and Muthen (www.statmodel.com). Specifically, MplusAutomation provides routines for creating related groups of models, running batches of models, and extracting and tabulating model parameters and fit statistics.
84 stars 46 forks source link

Bug extracting Tech4 when using WLSMV #105

Open martscht opened 5 years ago

martscht commented 5 years ago

Extracting the Tech4 currently does not seem to work with WLSMV (and I suspect other WLS-estimators). The returned Tech4 is empty even if there is a Tech4 in the Mplus output. The cause of this is inconsistent formatting of Tech4 in Mplus: when using ML estimators the output is grouped by matrices (going matrix, se, t, p for each matrix type) - when using WLSMV the output is gouped by content (going matrix, matrix, matrix for all types, then se, se, se for all types and so on). From the error-producing output with WLMSV:

` ESTIMATES DERIVED FROM THE MODEL

       ESTIMATED MEANS FOR THE LATENT VARIABLES
          RA1           RA2           RA3
          ________      ________      ________
            0.000        -0.006         0.074

       ESTIMATED COVARIANCE MATRIX FOR THE LATENT VARIABLES
          RA1           RA2           RA3
          ________      ________      ________

RA1 0.820 RA2 0.327 0.760 RA3 0.399 0.400 0.766

       ESTIMATED CORRELATION MATRIX FOR THE LATENT VARIABLES
          RA1           RA2           RA3
          ________      ________      ________

RA1 1.000 RA2 0.415 1.000 RA3 0.504 0.524 1.000 `

and the same model mit ML:

` ESTIMATES DERIVED FROM THE MODEL

       ESTIMATED MEANS FOR THE LATENT VARIABLES
          RA1           RA2           RA3
          ________      ________      ________
            0.000         0.015         0.014

       S.E. FOR ESTIMATED MEANS FOR THE LATENT VARIABLES
          RA1           RA2           RA3
          ________      ________      ________
            0.000         0.058         0.050

       EST./S.E. FOR ESTIMATED MEANS FOR THE LATENT VARIABLES
          RA1           RA2           RA3
          ________      ________      ________
            0.000         0.259         0.281

       TWO-TAILED P-VALUE FOR ESTIMATED MEANS FOR THE LATENT VARIABLES
          RA1           RA2           RA3
          ________      ________      ________
            1.000         0.795         0.779

       ESTIMATED COVARIANCE MATRIX FOR THE LATENT VARIABLES
          RA1           RA2           RA3
          ________      ________      ________

RA1 0.194 RA2 0.057 0.202 RA3 0.091 0.065 0.179 `

I'm running Mplus Version 8.0 on a Linux system.

sda030 commented 3 years ago

Rather than implementing a fix in MplusAutomation, should we not try to alert Mplus to this annoyance?