Cohens Conventions for Small, Medium, and Large Effects

These conventions should be used with caution. What is a small or even trivial effect in one context may be a large effect in another context. For example, Rosnow and Rosenthal (1 ! " discussed a 1 !! biomedical research study on the effects of ta#ing a small, daily dose of aspirin. $ach participant was instructed to ta#e one pill a day. For about half of the participants the pill was aspirin, for the others it was a placebo. The dependent variable was whether or not the participant had a heart attac# during the study. %n terms of a correlation coefficient, the si&e of the observed effect was r ' .()*. %n terms of percentage of variance explained, that is (.1+,. %n other contexts this might be considered a trivial effect, but it this context it was so large an effect that the researchers decided it was unethical to continue the study and the contacted all of the participants who were ta#ing the placebo and told them to start ta#ing aspirin every day.

Difference Between Two Means*

-i&e of effect small medium large d .+ .. .! , variance 1 / 1/

0ohen1s d is not influenced by the ratio of n1 to n2, but rpb and eta2s3uared are.

Pearson Correlation Coefficient

-i&e of effect small medium large .1 .) .. +. , variance 1

$ffect-i&e0onventions.doc

Contingenc Ta!le "nal sis

-i&e of effect small medium large w' .1 .) .. odds ratio4 1.* ).*.

4For a + x + table with both marginals distributed uniformly.

"#$%" Effect
-i&e of effect small medium large f .1 .+. .* , of variance 1 / 1*

Multi&le R2
-i&e of effect small medium large f2 .(+ .+. .* , of variance + 1) +/

'arl (uensch, East Carolina )niversit * +evised #ovem!er, ,--.* 5ore detail on these conventions and power Wuensch1s -tatistics 6essons