Do We Know a Vector From a Scalar? Why Measures of Association (Not Their Squares) Are Appropriate Indices of Effect



“Variance accounted for”—calculated by squaring one of the various measures of association—is the most common estimate of experimental effect or strength of association reported in communication studies. However, methodologists in other social science disciplines have made compelling cases that the statistic itself, not its square, is the appropriate index of shared variation. The basic principles and arguments for interpreting unsquared measures of effect are presented and the implications for the practice of communication theory and research are discussed.