M2 for latent regression and multidimensional models

[Balal Ezanloo]

Hi Phil

Just for saving the time see my below notes

1) M2 does not work for latent regression model

M2(LRM.r.p.2pl.out, na.rm = TRUE,calcNull = TRUE) 

Error in M2internal(obj, calcNull = FALSE, quadpts = quadpts, theta_lim = theta_lim,  : 

  Latent regression models not yet supported

2) empirical reliability does not work for latent regression model. there is no any 'empirical reliability in the output

LRM.r.p.score   <- fscores(LRM.r.p.2pl.out, method = "EAP", full.scores = TRUE, 

                       full.scores.SE=TRUE,returnER = T)

 head(LRM.r.p.score)

             F1     SE_F1

[1,] -1.3369287 0.5580695

[2,] -0.8654669 0.5962059

[3,] -1.1013687 0.5569815

[4,] -1.2882071 0.5868284

[5,] -0.5390334 0.5948275

[6,] -0.2679891 0.6156024

3) M2 doesn't work for two dimensional model

> M2(cfa.2d.out, na.rm = TRUE,calcNull = TRUE)

Error : The following items have only one response category and cannot be estimated: V12 V13 V14 V15 V16 V17 V18 V19 V20 V21 V22 V23 V24 V25 V26 V27 V28 V29 V30 V31 V32 V33 V34 V35 V36 

In addition: Warning message:

In cov2cor(cross/N - outer(colMeans(dat), colMeans(dat))) :

  diag(.) had 0 or NA entries; non-finite result is doubtful

Error: Null model did not converge or is not supported

> M2(cfa.2d.out, na.rm = TRUE,calcNull = F)

            M2   df p      RMSEA   RMSEA_5   RMSEA_95 SRMSR

stats 5302.729 1175 0 0.04427592 0.0430561 0.04547841   NaN

Warning message:

In cov2cor(cross/N - outer(colMeans(dat), colMeans(dat))) :

  diag(.) had 0 or NA entries; non-finite result is doubtful

4) can i use 'anova' function for compare the fit of different models, just because their data differ based on different treatment to NA?Any help? 

[Phil Chalmers]

Likely just the information criteria for now. Latent regression should be supported in the M2 statistic, but it isn't yet for reasons that I still have some work to do with latent regression models in mirt to make them more general.

Phil

On Tue, May 22, 2018 at 3:14 PM Balal Ezanloo <b.ez...@gmail.com> wrote:

Hi Phil

Can I use M2 for camparign the fit of 2pl, latent regression and multidimensional model? If not; which function of mirt can handle this problem (with exception of anova)? Any help?

--
You received this message because you are subscribed to the Google Groups "mirt-package" group.
To unsubscribe from this group and stop receiving emails from it, send an email to mirt-package...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.