A comparison of various strength of association measures commonly used in gerontological research.

An increased usage of measures of strength of association or effect size (ES) has been noted in the gerontological literature. Potential confusion could be generated by the fact that different researchers have reported different measures of ES from study to study. This paper involves a comparison of six ES measures commonly used by gerontological researchers as these measures relate to one another in both the analysis of variance and multiple regression models. Three other issues involving ES measures are discussed: (1) the ES of a contrast; (2) orthogonal and nonorthogonal designs; (3) partial Ess.