[Bug]: Pareto Chart of Standardized Effects changes when Uncoded Results shown

[bensagalovsky]

JASP Version

0.95.0

Commit ID

No response

JASP Module

Quality Control

What analysis are you seeing the problem on?

Analyse Design

What OS are you seeing the problem on?

Windows 10

Bug Description

When "Display results in coded units" is unchecked, the Pareto of Standardized Effects changes, and shows bars corresponding to the uncoded coefficients, while maintaining the t-critical value.

This is misleading. Even when unchecking "Display results in coded units", the Pareto should stay the same.

Expected Behaviour

When unchecking "Display results in coded units", the Pareto Chart of Standardized Effects should stay the same.

Let me take the opportunity to thank you for the Quality Control module, it is enabling me to use JASP as the default program for Six Sigma and other trainings, which would not be possible without it!

Steps to Reproduce

Just click and unclick "Display results in coded units" in the attached .jasp file, and see how the Pareto changes.

Analyse Design example.zip

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.

[tomtomme]

@bensagalovsky
thx for the report

@JTPetter
can you comment on this one? thx

[JTPetter]

@bensagalovsky Thanks for the report and happy to hear you find the module useful!

This is not a bug, but relates to the way the model is specified when you enter predictors as "continuous factors". If you treat predictors as continuous in the model, changing the scale by coding the predictors does indeed change the t-values, on which the Pareto chart is based. I believe this is correct behavior. If you treat the factors as discrete (which is probably what you want, since even though you are dealing with numeric variables, you have discrete levels of the variable), you will get the same results, whether the variables are coded or not. Hope this clarifies the issue? Let me know if you agree.

[tomtomme]

@bensagalovsky
can we close this one? Did JTPetters explanation help?

[bensagalovsky]

Yes indeed, thank you very much - and please excuse the delay in closing this subject.
Please excuse the annoyance, @JTPetter !