No missing values, but EFA with polychoric correlations says too many missings
Please explain this😭😭
The problem only occurs when the data are ordinal or nominal, not numeric (i.e. based on polychoric correlations)
Just as a sanity check, here's the data (SPSS file with labels):
Comments
This is in JASP 0.96.0, I don't know if it's new as of this new version
So it might be the way you have laid out your response options. Now I can't read French very well but I will give below the typical way of responding to the commonly used depression questionnaire:
Patient Health Questionnaire-9 (PHQ-9) - Mental Health Screening - National HIV Curriculum
I wonder if it is therefore inappropriate for your data to be nominal and they should in fact be ordinal or nominal? The problem is that I cannot tell the structure of your questionnaire you possibly get some more detail? I wonder if this might be a one case per variable type issue? If this doesn't help or isn't applicable I wonder if it might help to file a bug report? Sorry if none of this helps, just what came to mind.
Best,
Tarandeep
It may have to do with the fact that there is text in there as well. I will bring this to the attention of our expert.
Hi,
at first glance this looks more like a data reading issue than one with the analysis. How are other analyses working? In any case, can you open an issue on github.com/jasp-stats/jasp-issues and provide the data file there?
Julius
@julius it does look like it's because of the labels (as @EJ pointed out), because I tried with another dataset that also had value labels alongside numeric values and same error happened, but when I tried with a dataset that didn't have labels (only numeric values entered by hand in the data editor), the EFA with polychoric correlations ran fine.
I filed an issue here: https://github.com/jasp-stats/jasp-issues/issues/4129
+1 thanks