[Bug]: Population Effect Size and 95% Credible interval NOT showing when error: "Cannot plot the posterior"--possibly too concentrated near 0.

[Miss-Egg]

JASP Version

0.18.3

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 10

Bug Description

The error displayed: "Cannot plot the posterior - possibly too concentrated near 0."

I've ran a bayesian one samples t-test against 0. The alternative hypothesis is theta 0. Prior @ Default. I need to report a delta value and its 95% CI (even if the plot cannot be created on JASP).

I appreciate the time in addressing this issue! Just trying to get these into a table for my dissertation :).

example stim.csv

https://forum.cogsci.nl/discussion/comment/29423#Comment_29423

Expected Behaviour

Would like to have the values (population effect size and 95% credible interval) even if the plot cannot be displayed.

Steps to Reproduce

1. Click T-Test
2. Click Bayesian One sample t-test
3. Test value: type in 0
4. Alternative Hypothesis: not equal to test value selected
5. Bayes Factor: BF10
6. Plots: Select Prior and Posterior and Additional Info

Log (if any)

No response

Final Checklist

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

[tomtomme]

@Miss-Egg Confirmed. This is a duplicate of https://github.com/jasp-stats/jasp-issues/issues/1455 I closed that old one.

@vandenman So this is a really old bug. Can it be adressed for jasp 0.19?

Relevant remarks form the duplicate:

• Reproducible via Summary Stats independent t-test, enter n1=n2=20; with t = -12 > plotl works; t=-14 > plot fails
• the error message was added in https://github.com/jasp-stats/jasp-desktop/pull/3267
• The code that computes the posterior density returns infinity, code see https://github.com/jasp-stats/jasp-desktop/pull/1755 - It looks like the CI fails and that's why the plot can't be shown.

[Miss-Egg]

Hi @tomtomme and @vandenman, this bug is for the most recent release (0.18.3). I'm trying to get the effect size value at a minimum, and preferred, with its CI value. Because it is only displayed when a plot can be made, the value is not accessible via JASP.

[psychuser00]

I'm having this problem too in the bayesian one sample t-test: ""Cannot plot the posterior - possibly too concentrated near 0." I have a sample size of 10 participants, and when I try to use the default prior: Cauchy scale .707 against a test value of 0, it provides that error message. I've attached the data. bayesian one sample t-test error.csv.csv