Prof. McClelland's current research activity is focused primarily on statistical issues. In particular, he is an international authority on the statistical difficulties (i.e., power) of detecting interactions or moderator effects. He currently is collaborating with several colleagues here and at Duke to develop better graphical and statistical methods for explicating interaction effects. He has also done extensive work on the havoc created in statistical analyses by dichotomizing otherwise continuous variables, an all too common practice. Prof. McClelland also conducts research to understand graphical comprehension and to increase the quality and information density of statistical graphics.
judgment and decision making, statistics, moderator or interaction linear models