[Feature request]: Set missing data treatment automatically depending on type of data

[brianmsm]

JASP Version

0.18.3

Commit ID

No response

JASP Module

Factor, SEM

What analysis are you seeing the problem on?

No response

What OS are you seeing the problem on?

Windows 11

Bug Description

For some time now, JASP automatically detects when categorical variables are entered in the analysis and sets the appropriate estimator to be used to “WLSMV”, which is appropriate and moves towards a more straightforward and simple use of the software.

However, the first thing one notices when entering the variables or items in the software is the warning of

The model could not be estimated. Error message: Missing data handling 'FIML' is not supported for categorical data. Please select another method under 'Missing data handling' within the 'Advanced' options tab

This is because in the Factor and SEM module, the missing value handling method is automatically set to FIML when the estimator is set to “Auto” (default). This worked when the WLSMV estimator was not yet supported, but now it causes the first message of the analysis to be an error message.

Same situation with the SEM module:

categorical_test.zip

Expected Behaviour

When categorical data is present, the Missing data handling should be set to Listwise deletion

Steps to Reproduce

1. With any database with categorical variables, open the analysis module and add the items in a factor.

Log (if any)

No response

Final Checklist

• [X] I have included a screenshot showcasing the issue, if possible.
• [X] I have included a JASP file (zipped) or data file that causes the crash/bug, if applicable.
• [X] I have accurately described the bug, and steps to reproduce it.

[juliuspfadt]

good point. I will label it as a feature request for now though.

[juliuspfadt]

@boutinb is this possible within qml? set an option based on the measurement level of the variable(s)?