The Effect Size Statistic

An increasing array of health status instruments provides challenges to the researcher in selecting the most appropriate measure in health outcomes studies. The responsiveness of a measure is its ability to identify small but important changes over time or following interventions. For health outcomes studies, responsiveness is an important psychometric requirement to guide instrument selection and it can be evaluated using the effect size statistic. The calculation and meaning of effect sizes are described here. Effect size considerations can also contribute to other topics of importance to health researchers. Issues of power and statistical and clinical significance are discussed.