Three aspects of an empirical effect: statistical, theoretical, and practical aspect
A transparent evaluation of an empirical effect’s relevance is based on the size of effect (statistical aspect), a theoretical construct’s ability to adequately predict the effect (theoretical aspect), and the effect’s practical utility (practical aspect). In behavioral science publications, however, all three aspects are often found conflated. Already if only the practical aspect is evaluated independently of the other two aspects, disagreements about the effect’s relevance turn out to be resolvable. And, if also the statistical aspect is evaluated independently of the theoretical aspect, then the ‘smallest effect of interest’ turns out to be much larger when predicting an effect (statistical aspect) as opposed to explaining it (theoretical aspect). Crucially, behavioral science publications today typically report either small, homogenous empirical effects or large, heterogeneous ones. This pattern greatly impairs the prospects for theory construction in behavioral science, because an empirically adequate theoretical construct would have to predict a larger and more homogenous empirical effect than can be observed.