[Bug]: Classic Process Models uninformative errors when categorical predictor levels have spaces

[bnevicka]

JASP Version

0.95.0

Commit ID

No response

JASP Module

Process

What analysis are you seeing the problem on?

I tested Model 1 and Model 4 and neither of them ran

What OS are you seeing the problem on?

macOS Silicon

Bug Description

I was testing a simple moderation model (Model 1) with a categorical predictor and continuous moderator and the model did not run and reported the following error:

_Error in str2lang(paste(vSplit, collapse = '*')): :1:29: unexpected symbol
1: emregemotion regulation
^

Stack trace
tryCatchOne(expr, names, parentenv, handlers[[1]])

doTryCatch(return(expr), name, parentenv, handler)

withCallingHandlers(expr = analysis(jaspResults = jaspResults, dataset = dataset, options = options), error = .addStackTrace)

analysis(jaspResults = jaspResults, dataset = dataset, options = options)

.procModProbes(jaspResults, dataset, options)

.procModProbesSingleModel(modelsContainer[[modelName]], dataset, options)

.procMeanCenter(container[['graph']]$object, dataset, options)

eval(str2lang(paste(vSplit, collapse = '*')), envir = dataset)

str2lang(paste(vSplit, collapse = '*'))_

I then ran model 4 (simple mediation) and this model also did not provide results and only the following error:

Estimation failed because of an internal error.

Expected Behaviour

This is not an issue with the datafile because I tested both of these models in SPSS with the Process macro and they both worked and provided results.

Steps to Reproduce

1. I added in the relevant variables
2. I then specified the Hayes model as either model 1 or model 4
3. I then added the relevant variables for the model (i.e., predictor and then moderator (W) or mediator)

Log (if any)

No response

More Debug Information

No response

Final Checklist

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

[JohnnyDoorn]

@maltelueken @ThijsVroegh can you please take a look asap? this seems like a pretty severe bug

@bnevicka could you please include the data file and configuration (or just the jasp file that includes the error) here? They can be uploaded by creating a zip file and uploading to github

[EJWagenmakers]

Also, @bnevicka , the nature of the error message suggests that perhaps there is something unexpected going on with the variable names. Did you check that the model still does not run if you choose variable names like "bla1", "bla2" etc.?

[bnevicka]

@EJWagenmakers , I just checked and the same error message remains even when I change the variable names to bla1, bla2 etc.

[bnevicka]

@JohnnyDoorn , I emailed the jasp file with the three models and error messages

[JohnnyDoorn]

Thanks @bnevicka! The issue is related to the levels having spaces in them - removing the spaces (e.g., "control" and "training") will solve the problem.

(my initial response was to change the variable type, so that the 0 and 1's will be analyzes rather than the level names in the subsequent analyses; that also solves the problem but didn't get to the core of it)

[bnevicka]

@JohnnyDoorn , thanks very much! That's helpful to know.