search

philchalmers / mirt

Multidimensional item response theory
https://philchalmers.github.io/mirt/
199 stars 75 forks source link

supporting broom, df.residual method and pooling model estimation (mice::pool) from multiple imputation #187

Closed tingjhenjiang closed 3 years ago

tingjhenjiang commented 4 years ago

Hello, I would like to apply mirt analysis on multiple imputated survey datasets. After I got mirt models for differently imputed datasets, I tried to pool the results by R package pool, but in vein. There is a discussion at the github page of R package mice. Can you add some S3methods or S4methods of broom::glance(), broom:tidy(), df.residual() to a mirt model ( e.g., class SingleGroupClass )? You might check tidymirt project and the readme page of mice::pool().

Thank you!

seonghobae commented 4 years ago

Unfortunately, mirt already has a multiple imputation function.

tingjhenjiang commented 4 years ago

Unfortunately, mirt already has a multiple imputation function.

hello, Do you mean mirt::averageMI ? I think it is not straightforward according to the given example.

philchalmers commented 4 years ago

Thanks for the suggestion. Do you have suggestions as to what coding work would be required in order for mirt to behave nicely with broom? I'm not sure what would be required to implement an S3 method for broom to work with (I don't have interest in adding broom as a dependency). Thanks.