APA Dictionary of Psychology

a phenomenon that can occur when data from two or more studies are merged, giving results that differ from those of either study individually. For example, two studies, each showing a correlation of .00 between two variables, x and y, may show a strong positive correlation between variables x and y when the data are merged. This paradoxical effect could occur if the mean values of each variable in one study are both substantially different (e.g., lower) than their mean values in the second study. [Edward H. Simpson (1922–  ), British statistician]