I'm hoping to report effect sizes and their corresponding magnitudes (i.e., small, medium, large) for HLM analyses run using lme4 in R. Initially, I was planning to report pseudo-R^2 (MuMIn package), however I wasn't able to output this statistic for my binomial logistic HLM models. I've also read that Cohen's w can be used to evaluate effect size for incremental steps in hierarchical models. From the literature, there doesn't appear to be consensus regarding measures of effect size in mixed models and it doesn't seem as though many people use Cohen's w. Any advice would be appreciated!


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.