When the effect of one thing depends on another: Not a reason for complete pessimism

Abstract

Consider the joint effects of several factors on some quantitative outcome variable y. If it is found that the effect ofone depends upon the others, there is said to be 'interaction' between the factors. This article will argue that this may be due to a non-linear dependence ofy on a single input, which itself is simply the sum of the effects of the variables being studied. One thing that stimulated this article was the rather pessimistic body ofliterature that emphasises the number of variables in international relations, the variety of their combinations, the potential interactions between them, and the unpredictability of the result. Much of this makes reference to complexity and chaos theory. By proposing how interaction might be modelled, this pessimism may be countered. Two examples of this are Kerbel (2004) and Geeraerts (1998).