Cautionary note on the use of omega squared to evaluate the effectiveness of behavioral treatments

Estimating the magnitude of treatment effects has been recommended as a solution to the problems associated with conventional hypothesis testing. In comparison to tradition statistical tests of treatment effectiveness, omega squared (ω2) and related magnitude of effect statistics provide a graduated rather than a dichotomous judgmental aid, index the strength of the relationship between treatment and outcome, and are unaffected by aspects of statistical power related to sample size. Unfortunately, these correlational statistics also have characteristics that limit their interpretative value.