Question regarding the impute argument in older mirt version

[Oskar Engels]

Hello everyone,

I am currently working through an example from the "Handbook of Educational Measurement and Psychometrics Using R" by Desjardins and Bulut. The example pertains to item fit of a Rasch model using the mirt package.

Their R code was:

rasch_itemfit <- itemfit(rasch_fit,
fit_stats = c("S_X2", "G2"),
impute = 10)
head(rasch_itemfit)

This results in:

item        G2      df.G2   p.G2    S_X2    df.S_X2  p.S_X2
1 reason.4   10.888   9     0.344   20.597  12       0.057
2 reason.16  9.007    9     0.443   10.802  12       0.546
3 reason.17  15.409   9     0.131   19.920  12       0.069
4 reason.19  14.018   9     0.166   26.702  12       0.009
5 letter.7   5.458    9     0.788   12.490  12       0.407
6 letter.33  14.195   9     0.163   27.096  13       0.012

However, when I run the same code, I encounter the following error message:

Only X2, G2, PV_Q1, PV_Q1*, infit, X2*, and X2*_df can be computed with missing data.
Pass na.rm=TRUE to remove missing data row-wise

I understand that something has changed regarding the impute argument in the mirt package since the book was published. How can I recreate their results with the current version of the mirt package? Should I use imputeMissing or a different function?

Cheers,

Oskar

[Phil Chalmers]

I would impute the data based on other methods, like multiple imputation from other dedicated packages. Otherwise, you'll have to use the na.rm inputs instead as the previous imputation approach was rather suboptimal. HTH.

Phil

--
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.
To view this discussion on the web visit https://groups.google.com/d/msgid/mirt-package/168336c2-f40e-4508-a218-007c33a97e5cn%40googlegroups.com.