Clinical Information Regarding Effect Size

The translation of research findings into clinical practice requires that clinicians and policy makers be competent in understanding pivotal clinical research. Comparing two groups using an effect size or standardized mean difference (e.g. Cohen's d, Hedges' g) has become increasingly common in the past few years. Effect side is the ratio of the difference between groups divided by the calculated standard deviation. This is akin to a signal to noise ratio; the higher the value the more noticeable the signal compared to the noise.

 

Cohen (1969, 1988) hesitantly offered the following thresholds for clinically interpreting effect size calculations and provides the following common examples:

Effect Size Clinical Interpretation Lay Example
0 No effect Complete overlap; no difference in means
0.2 Small clinical effect Difference in heights of 15 and 16 year old girls
0.5 Medium clinical effect Difference in heights of 14 and 18 year old girls
0.8 Large clinical effect Difference in heights of 13 and 18 year old girls, or difference in IQ between PhDs and ’typical college freshmen’

Interpreting the clinical significance of an effect size is not intuitive. The effect size thresholds, offered by Professor Cohen, of 0.2, 0.5, and 0.8 have been extensively cited as relating to small, medium, and large clinical effects, but are not without criticism. What is often overlooked is what else Professor Cohen said: "The terms 'small', 'medium', and 'large' are relative... to each other... the definitions are arbitrary... these proposed conventions were set forth throughout with much diffidence, qualifications, and invitations not to employ them if possible.” “The values chosen had no more reliable a basis than my own intuition."

 

An effect size needs to be interpreted with consideration of the observed difference between means in context of illness severity or burden, measured variance, and other factors. For a more complete list see What is the ESI: Interpreting Effect Sizes.