search

jasp-stats / jasp-issues

This repository is solely meant for reporting of bugs, feature requests and other issues in JASP.
59 stars 29 forks source link

[Feature Request]: Robust t-test and ANOVA #2769

Open JaspBoy opened 4 months ago

JaspBoy commented 4 months ago

Description

Robust t-test and ANOVA for violation of assumptions of general linear model

Purpose

To conduct sensitivity analysis which most on statistics literature now recommend

Use-case

No response

Is your feature request related to a problem?

No response

Is your feature request related to a JASP module?

Robust T-Tests, T-Tests

Describe the solution you would like

Would like implementation of robust t-tests and ANOVA in JASP

Describe alternatives that you have considered

JAMOVI program

Additional context

Images from JAMOVI

slika slika slika

EJWagenmakers commented 4 months ago

I think this functionality would in general be great to have. Andy Field is also a fan. And these methods are in general the bread and butter of the empirical sciences. I believe @JohnnyDoorn had an interest to look at this...

JaspBoy commented 4 months ago

I think this functionality would in general be great to have. Andy Field is also a fan. And these methods are in general the bread and butter of the empirical sciences. I believe @JohnnyDoorn had an interest to look at this...

Sounds great! Also a fan of Andy Fields books. Can't wait for the book "Discovering statistics using JASP". I was so thrilled to see this collaboration.

Any predictions when could robust modules be introduced to JASP? :)

TarandeepKang commented 4 months ago

Hi all I would have to agree that robust testing would be an important addition! Regarding t-tests this would be a useful read, I think:

Karch, J. D. (2021). Choosing Between the Two-sample T Test and its Alternatives: A Practical Guideline. OSF. https://doi.org/10.31234/osf.io/ye2d4

and as for range of other robust tests, this i could be useful? #2305

In particular when it comes to Brunner Munzel as an alternative to the Mann-Whitney U.

Patrick Mair has a set of functions for other robust procedures:

https://link.springer.com/article/10.3758/s13428-019-01246-w

So this might indeed be your preferred "one-stop shop"?

There is however, reason to go beyond robust testing to permutation-based tests as suggested #454 #2305 Best,

Tarandeep

JohnnyDoorn commented 4 months ago

I am indeed planning to implement Rand Wilcox' WRS package in JASP - it will not be in the upcoming release, but hopefully the one after (so probably spring 2025).

JaspBoy commented 4 months ago

I am indeed planning to implement Rand Wilcox' WRS package in JASP - it will not be in the upcoming release, but hopefully the one after (so probably spring 2025).

Looking forward to it and thank you for your hard work!

Kind regards Ivan

TarandeepKang commented 4 months ago

Hi @JohnnyDoorn I just want to clarify did you mean that you wanted to consider using the WRS package https://github.com/nicebread/WRS or WRS2 which is the one you have linked to above? Although they have similar names, it looks as if WRS actually has a wider variety of features and is more up-to-date. See for example this paper:

Wilcox, R. R., & Rousselet, G. A. (2023). An Updated Guide to Robust Statistical Methods in Neuroscience. Current Protocols, 3(3), e719. https://doi.org/10.1002/cpz1.719

Best,

Tarandeep

JohnnyDoorn commented 4 months ago

Hi @TarandeepKang,

Thanks for the suggestion - the WRS package indeed seems more accessible/open, also because I cannot seem to find the WRS2 Github repository.