Document "Are there discrepancies in the predictors (with respect to numerical congruence)?"

[nicebread]

Document this output in the help file of the RSA function.

Hi! I was looking for some information on exactly this, but can't find anything in the documentation. When running the RSA function, in the output from summary() I get percentages like 2%, 6% etc. for each category, seldom adding to more than 15-20%. Is there any information available on how these should be interpreted, thresholds, what can be done if they are too low etc.? Thanks for an awesome package, btw!

[nicebread]

These three categories should add up to 100%. Do you have the most current version (we once fixed a bug in the computation of these numbers)? If yes, then we have to investigate it.

Thanks for the quick reply! I'm using the newest version on CRAN (0.10.6). It's a fairly large dataset, but with a good deal of missing. That is to say, the data frame I'm using has around 2000 rows, but the relevant variables for this analysis only have data in around 500 of these. Could this somehow be related? I figured that rows with missing were deleted listwise in the regression anyway so it didn't matter for the results, but could it affect how the percentages are counted?

[nicebread]

OK. The easiest way would be (if possible), if you send me the data set (only relevant variables) to my email address felix.schoenbrodt@psy.lmu.de