2
$\begingroup$

In the SPSS output when using the MIXED procedure (analyse -> mixed models -> linear) you only get the F value and the degrees of freedom. I found some formulas for estimating partial eta squared using only F and dfs (e. g. in the paper "Calculating and reporting effect sizes to facilitate cumulative science: a practical primer for t-tests and ANOVAs" by Daniël Lakens; https://doi.org/10.3389/fpsyg.2013.00863).

It says you can only use the formulas if you got no covariates and no measured factors. The problem is I got several covariates in my model with a within and a between subjects factor and a few interactions (see below). How can I calculate partial eta squared (or SSerror and SSeffect)?.

Here is the Syntax:

MIXED Y BY X1 X2 WITH COV1 COV2 COV3

/FIXED=X1 X2 COV1 COV2 COV3 X1*X2 X1*COV1 X1*COV2 | NOINT SSTYPE(3)

/REPEATED=X2 | SUBJECT(id) COVTYPE(UN).

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.