Abstract
Experiments in cognitive psychology usually return two dependent variables: the percentage of errors and the reaction time of the correct responses. Townsend and Ashby (1978, 1983) proposed the inverse efficiency score (IES) as a way to combine both measures and, hence, to provide a better summary of the findings. In this article we examine the usefulness of IES by applying it to existing datasets. Although IES does give a better summary of the findings in some cases, mostly the variance of the measure is increased to such an extent that it becomes less interesting. Against our initial hopes, we have to conclude that it is not a good idea to limit the statistical analyses to IES without further checking the data.