Effect size for two-way ordinal ANOVA with cumulative link model

Asked Mar 08 '23 at 13:15

Active Mar 09 '23 at 09:50

Viewed 37 times

I am looking for an effect size measure that is suitable for a two-way (2x2) ordinal ANOVA with cumulative link model.

I have seen the functions rank_eta_squared() and rank_epsilon_squared() but in my understanding and after trying, they only work for one-factorial designs. Is there anything comparable for two-factorial designs?

What I have done so far is estimating my model with clm() function, displaying wald chi square statistics and tests for main effects and interaction with anova() function and then calculating Cramers V with sqrt(chi²/n*df) while df is always 1. Is this an acceptable approach?

Thanks!

Edit Example:

library(ordinal)
n <- 100
dat <- data.frame(IV1=sample(LETTERS[1:2], n, replace=TRUE),
                  IV2=sample(LETTERS[3:4], n, replace=TRUE),
                  DV=factor(sample(1:6, n, replace=TRUE), ordered = TRUE))

mod <- clm(DV ~ IV1 + IV2 + IV1:IV2, data = dat)
anova(mod)
sqrt(anova(mod)[2]/nrow(dat)) # Cramers V

edited Mar 09 '23 at 09:50

asked Mar 08 '23 at 13:15

Adrian

Can you make a [minimal, reproducible example](https://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example) of what you've tried so far? – jrcalabrese Mar 08 '23 at 14:04
Sure, I edited an example. – Adrian Mar 08 '23 at 15:23

Effect size for two-way ordinal ANOVA with cumulative link model

0 Answers0