[Bug]: Wrong term for Confidence Intervall

[Andreas5703]

JASP Version

0.95.0

Commit ID

No response

JASP Module

T-Tests

What analysis are you seeing the problem on?

Bayesian One Sample T-Test

What OS are you seeing the problem on?

Windows 11

Bug Description

At "Descriptives" in Bayesian One Sample T-Test there is mentioned a 95 % Credible Intervall.
Even if the Prior will be modified in wide rages this intervall doen't change: It is not a Credible Intervall it is exacly the frequentist Confidence Intervall!

Expected Behaviour

It would be very helpful to get Credible Intervalls for Bayesian Estimations.

Steps to Reproduce

Carry out Bayesian One Sample T-Test
Show "Descriptives"

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.

[github-actions]

@Andreas5703, thanks for taking the time to create this issue. If possible (and applicable), please upload to the issue website (#3573, attaching to an email does not work) a screenshot showcasing the problem, and/or a compressed (zipped) .jasp file or the data file that causes the issue. If you would prefer not to make your data publicly available, you can send your file(s) directly to us, issues@jasp-stats.org

[EJWagenmakers]

Thanks for raising this issue. This is meant as a feature, not a bug. The 95% CI is from flat priors (and hence happens to correspond to the classical CI). The reason for that is that the plot is descriptive. The help file should say something about this! I checked, and the only thing that the help file now says is this:

"Descriptives plots
Credible interval: Default is 95%."

This is clearly insufficient! So we need to choose between two options:
(a) changing the output, and showing the Bayesian CI that corresponds to the specified prior;
(b) retaining the output, and at a minimum specify in the help file that this CI is based on flat priors. It is strange -- I thought we made that adjustment already. Maybe it was for ANOVA, I am checking this now....

OK, for the Bayesian ANOVA the help file says:
"Display:
Credible interval: When this option is selected, the plot will contain central credible intervals. By default this is set to 95%. This can be changed into the desired percentage."

But it is not clear whether the interval is the wide one, based on the the vague priors.
The thing is, if we do Descriptives for ANOVA, should we also base them on the non-flat priors and the complete model. If so, what model?

So because of the problem posed by ANOVA, I am still in favor of using flat priors for the descriptive plots, but we should clearly state what they are.

@vandenman can you do this? I would make the help file text consistent across all three t-tests and the Bayesian ANOVA's; maybe it can say something like:
"Credible interval: When this option is selected, the plot will contain central credible intervals based on flat prior distributions; the results will coincide with the frequentist confidence interval. The plot uses flat prior distributions because it shows a descriptive result that should not change depending on the details of the statistical model. The default percentage is set to 95%."

Is there any other way to highlight this? Some footnote text somewhere? A mouseover on "Credible interval"?