Journalists covering medical research need to quantify the results of a study, which means they rely on effect size—the magnitude of difference between groups. If a medication in an randomized controlled trial has a large effect size — if 80% of patients respond extremely well to it — then it has solid clinical significance. A smaller effect size — perhaps only 5% of patients have a partial response — likely won’t pass muster.

There aren’t hard and fast rules for how much an effect size is needed to determine an intervention’s or exposure’s clinical significance, but there are very clearly defined ways to measure effect size. The most common and familiar of these include risk ratio, relative risk, hazard ratio and odds ratio. To better understand these, check out the tip sheet provided by Journalist’s Resource on what effect sizes mean — and don’t mean — and how to report responsibly on them.