January 4, 2017

Effect size

The effect size is a statistic which estimates the magnitude of an effect (e.g. mean difference, regression coefficient, Cohen's d, correlation coefficient). It can be used as a relevant interpretation of an estimated magnitude (weak, moderate, or strong effect). An effect size can be displayed as both an unstandardized and a standardized value.

If the original units of measurements are meaningful, the presentation of unstandardized effect statistics is preferable over that of standardized effect statistics. Meta-analysis benefit from knowing the original units, as difference in measured quantities regarding the same subject could result in differences in standardized effect size estimations, which in turn bias the outcome of a meta-analysis.

Mean difference (MD)

Standardized effect statistics are always calculable if sample size and standard deviation are given along with unstandardized effect statistics.

Cohen's d

1. Nakagawa S, Cuthill I. Effect size, confidence interval statistical significance: a practical guide for biologists. Biol Rev 2007;82:591-605.