[Bug]: Bayesian RM anova crashes with many factors

mdechas65 commented 2 years ago

JASP Version

0.16.2

Commit ID

No response

JASP Module

jaspAnova

What analysis are you seeing the problem on?

Bayesian Repeated measures ANOVA

What OS are you seeing the problem on?

Windows 10

Bug Description

I have been trying to run a Bayesian repeated measures ANOVA with 5 factors (4 within - all with 2 levels, and 1 between with 2 levels). I have tried to run this analysis around 20 times on more and more powerful machines but the analysis never finishes (although it starts to run for a few hours) - either I see only a blank JASP screen, or JASP shuts down completely. Please note that I can run the same analysis with no problems when there are fewer factors (i.e., 3 within and 1 between, all with 2 levels). Just wondering why I cannot run this analysis with 5 factors?

Expected Behaviour

I should be able to see a results table

Steps to Reproduce

Bayesian repeated measures ANOVA
Four within-S factor (2 levels each); 1 between-S factor (2 levels)
Options - BF10; Table - effects across matched models; Order - compare to null model ...

Log (if any)

No response

Final Checklist

[X] I have included a screenshot showcasing the issue, if any.
[X] I have included a JASP file that causes the crash/bug, if any.
[X] I have accurately described the bug, and steps to reproduce it.

juliuspfadt commented 2 years ago

HI @mdechas65, thanks for reporting this. One of the maintainers might have an idea why this happens. Could you provide a .jasp file for this bug, or at least a data file?

vandenman commented 2 years ago

Hi @mdechas65 can you provide the model specification? By default JASP considers all possible interactions between all within and between factors. These are highlighted in blue in the screenshot below:

If you are not interested in these interactions you can remove them from the model specification. That implies these will not be considered and then the analysis should run much faster.

mdechas65 commented 2 years ago

Hi there - many thanks for getting back to me. I have attached the data set - the first column can be ignored, the 2nd is the between-S factor of group, then 16 columns of data. The within-S factors are as follows (from slowest to quickest cycling): hemisphere (L, R); region (peri, hippo); task (s, d); item (IPIP_IPRP, NPNP). Hope this makes sense? - many thanks again, Marianne

From: Julius Pfadt @.> Sent: Monday, May 23, 2022 2:22 PM To: jasp-stats/jasp-issues @.> Cc: DeChastelaine, Marianne @.>; Mention @.> Subject: Re: [jasp-stats/jasp-issues] [Bug]: Bayesian RM anova crashes with many factors (Issue #1715)

HI @mdechas65https://nam02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fmdechas65&data=05%7C01%7Cmdechast%40utdallas.edu%7C12dc6b547ea347ed43ca08da3cf17d44%7C8d281d1d9c4d4bf7b16e032d15de9f6c%7C0%7C0%7C637889305134192795%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=ZGDKiBvACqFXZzlHXVCZkVFA4pQjPqevIWT7MOQBCa4%3D&reserved=0, thanks for reporting this. One of the maintainers might have an idea why this happens. Could you provide a .jasp file for this bug, or at least a data file?

- Reply to this email directly, view it on GitHubhttps://nam02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fjasp-stats%2Fjasp-issues%2Fissues%2F1715%23issuecomment-1135052044&data=05%7C01%7Cmdechast%40utdallas.edu%7C12dc6b547ea347ed43ca08da3cf17d44%7C8d281d1d9c4d4bf7b16e032d15de9f6c%7C0%7C0%7C637889305134192795%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=OxzCeTGqmpynsP7wJ5cZsTx5zD6LxzQGvk2HbIqnCFk%3D&reserved=0, or unsubscribehttps://nam02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FAHVHRT72HNPEVUQ2M6TPPITVLPLE5ANCNFSM5WWVQATA&data=05%7C01%7Cmdechast%40utdallas.edu%7C12dc6b547ea347ed43ca08da3cf17d44%7C8d281d1d9c4d4bf7b16e032d15de9f6c%7C0%7C0%7C637889305134192795%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=7d2Lb%2Fv4FV4OUdYn5oUlf2VW6mFPBHscQM42Lz6x8hA%3D&reserved=0. You are receiving this because you were mentioned.Message ID: @.**@.>>

mdechas65 commented 2 years ago

Many thanks for your comment Vandenman - I am right now running this with fewer interactions specified in the model and have some results.

mdechas65 commented 1 year ago

Hi there - also just to let you know the specifications I used. This was a repeated measures ANOVA with 4 within-S factors (hemi, region, task, item) and 1 between-S factor (group). I am using Bayes Factor BF10, order - compare to null model, and Tables - effects across matched models - thanks again, Marianne

From: Julius Pfadt @.> Sent: Monday, May 23, 2022 2:22 PM To: jasp-stats/jasp-issues @.> Cc: DeChastelaine, Marianne @.>; Mention @.> Subject: Re: [jasp-stats/jasp-issues] [Bug]: Bayesian RM anova crashes with many factors (Issue #1715)

HI @mdechas65https://nam02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fmdechas65&data=05%7C01%7Cmdechast%40utdallas.edu%7C12dc6b547ea347ed43ca08da3cf17d44%7C8d281d1d9c4d4bf7b16e032d15de9f6c%7C0%7C0%7C637889305134192795%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=ZGDKiBvACqFXZzlHXVCZkVFA4pQjPqevIWT7MOQBCa4%3D&reserved=0, thanks for reporting this. One of the maintainers might have an idea why this happens. Could you provide a .jasp file for this bug, or at least a data file?

- Reply to this email directly, view it on GitHubhttps://nam02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fjasp-stats%2Fjasp-issues%2Fissues%2F1715%23issuecomment-1135052044&data=05%7C01%7Cmdechast%40utdallas.edu%7C12dc6b547ea347ed43ca08da3cf17d44%7C8d281d1d9c4d4bf7b16e032d15de9f6c%7C0%7C0%7C637889305134192795%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=OxzCeTGqmpynsP7wJ5cZsTx5zD6LxzQGvk2HbIqnCFk%3D&reserved=0, or unsubscribehttps://nam02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FAHVHRT72HNPEVUQ2M6TPPITVLPLE5ANCNFSM5WWVQATA&data=05%7C01%7Cmdechast%40utdallas.edu%7C12dc6b547ea347ed43ca08da3cf17d44%7C8d281d1d9c4d4bf7b16e032d15de9f6c%7C0%7C0%7C637889305134192795%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=7d2Lb%2Fv4FV4OUdYn5oUlf2VW6mFPBHscQM42Lz6x8hA%3D&reserved=0. You are receiving this because you were mentioned.Message ID: @.**@.>>

tomtomme commented 7 months ago

@mdechas65 Since you resolved the problem I am closing this one! Feel free to reopen, if I am mistaken!

jasp-stats / jasp-issues