Measures of effect size in the reporting of rehabilitation research

Abstract

Effect size measures were computed for a sample of 100 research studies published in four leading rehabilitation journals. The effect size measures were categorized as small, medium, or large according to established criteria. Analysis of the effect size measures revealed that the average treatment effect was in the medium range (0.69). Further data analysis exposed a statistically significant difference in mean effect size values across articles contained in the four rehabilitation journals. A regression analysis indicated that characteristics of the study design were not significant predictors of effect size values in the 100 studies. The interpretation of effect size measures in relation to statistical significance testing is discussed and the argument made that reporting measures of effect size will assist researchers in determining the sensitivity of experimental manipulations in rehabilitation research.