Conceptual and analytic issues in crowding research

Three major mechanisms have been posited to explain the adverse effects of crowding on human health and behavior: behavioral constraint, diminished control, and overload/arousal. Theoretically, one or more of these underlying mechanisms is activated by high density. The operation of the mechanism(s), in turn, then functions as a mediator accounting for the ill effects of high density. Unfortunately, crowding researchers have often incompletely tested these hypothetical models. Frequent tests of the respective models have entailed orthogonally crossing the hypothetical, explanatory variable (e.g. control) representing the intervening mechanisms with density. The test of the model has then focused on the interaction of the explanatory variable (e.g. control) with density. Here we show that an interaction analysis specifies the explanatory mechanism as a moderator and is not a direct test of the respective models of crowding. Alternatively, some researchers have examined the direct effects of density on the explanatory variable (e.g. arousal) but failed to also evaluate the intervening effects of the mediating mechanisms between density and some index of health or well being. We show how prior crowding research, as well as more complex crowding models, can be more clearly conceptualized and more thoroughly tested utilizing the moderator-mediator variable distinction.