Statistical Question

Effect sizes

BMJ 2012; 345 doi: http://dx.doi.org/10.1136/bmj.e7370 (Published 5 November 2012)
Cite this as: BMJ 2012;345:e7370

Recent rapid responses

Rapid responses are electronic letters to the editor. They enable our users to debate issues raised in articles published on bmj.com. Although a selection of rapid responses will be included as edited readers' letters in the weekly print issue of the BMJ, their first appearance online means that they are published articles. If you need the url (web address) of an individual response, perhaps for citation purposes, simply click on the response headline and copy the url from the browser window.

Displaying 1-1 out of 1 published

14 November 2012

Dear Editor;

I am glad to see BMJ again taking the lead on understanding research findings. Effect size is a very poorly understood but increasingly common expression of "difference" between groups. I am a clinician and have been teaching how to understand effect size expressions to clinical students for several years. Upon their urging I developed a website with an effect size illustrator, insights into the meaning and interpretation of effect size, and numerous clinical examples, to help expand the clinical translation of effect size expressions.

This new website can be found at: http://esi.medicine.dal.ca. The effect size illustrator (hence "esi") is dynamic. The user can generate comparative images to better understand the "size of the difference" between groups. These images can take various shapes (normal distributions vs. my "ornament"). Titles, axes labels, and units can be added (when using option 1) when calculating and illustrating an effect size. If wanting to better understand an effect size that is already calculated (e.g., from a meta analysis), option 2 can be used to generate an illustration and interpretation.

My hope is that users of the illustrator and website will become increasingly more comfortable with interpreting comparisons using familiar (blood pressure) and unfamiliar (many specialty specific scales) continuous variable measures. In healthcare, I associate effect size measures with our management of "misery" (e.g., reduction in pain or depression) in contrast to our management of "risk" (therapy decisions based on NNTs in effort to prevent future events).

Competing interests: None

David M Gardner, Professor of Psychiatry & Pharmacy

Dalhousie University, AJLB Rm 7517, 5909 Veterans' Memorial Lane, Halifax, NS, Canada, B3H 2E2

Click to like:

THIS WEEK'S POLL