Generalized eta squared in mixed repeated measures ANOVA

[snow]

Hello,

I've ran into an issue when using jamovi for calculating a mixed repeated measures ANOVA.
The design includes two within factors (both with two levels) and one between subjects factor (also two levels).

I encounter an issue when calculating the generalized eta squared for the between-subjects effects.
In my specific example I obtain a unreasonably huge generalized eta squared. In this case, the between subjects effect is close to significance, it is a small effect, the partial eta squared is 0.12 and the generalized eta 2 is 0.41.
Is it possible that the generalized eta squared can be about four times the partial eta squared?

Coincidently it is also the same value as the generalized eta squared of my first within-subjects main effect (which indeed is a large main effect so this effect size would make sense here).

When comparing the results of my jamovi analyses with SPSS and jasp, the stats are 99% identical - however, the generalized eta squared is different between jasp and jamovi. Thus, I suspect a possible issue here.

Any help would really be appreciated.

Best
Martin

[jonathon]

hi,

can you attach a .omv file?

with thanks

[snow]

Thanks for the fast reply!
I've attached an .omv file.

Cheers
Martin

[jonathon]

ooh, you might not have quite attached it correctly. it doesn't seem to be here.

jonathon

[snow]

sorry - I was too slow I think. now it should (hopefully) be there.

Martin

[jonathon]

ah, you're quite right. good catch.

we'll push out a new version shortly.

here's a version of your file with the correct results:

Exemplary_MxdRepMeasuresANOVA_2x2x2_VidCatxTask_View_red-fixed.omv

(90.24 KiB) Downloaded 372 times

many thanks

jonathon

[snow]

Thanks to you for the super-fast help!

cheers
martin