[Skip to Navigation]
July 2011

Arguments for and Against Standardized Mean Differences (Effect Sizes)

Arch Pediatr Adolesc Med. 2011;165(7):592-596. doi:10.1001/archpediatrics.2011.97

Researchers often estimate an association to quantify a possibly causal relationship between an exposure (a treatment, habit, or experience) and an outcome (death, illness, or a continuous measurement). Statistics used to quantify associations include ratios or differences in risks, rates, or means. If the estimated association is causal, these statistics describe the effect of exposure on the outcome. The effect for a population is the average outcome if all were exposed (or at some exposure level) compared with the average outcome if all were not exposed (or at a different level of exposure).1,2

Add or change institution