Abstract
Recently, human ecological research has made increasing use of causal modeling in the attempt to understand aggregate-level social processes. This approach has great appeal because it helps make hypotheses explicit, provides a convenient way of structuring the application of statistical controls, allows the representation of complexity in the relations among variables, and, if strong assumptions are plausible, makes possible weak probes of causal hypotheses using only correlational evidence. The present report critically ex amines the usefulness of the path analytic approach in human ecology by exploring the rate of high school crimes reported to the police and providing some tests of the plausibility of the assumptions underlying the research strategy. It is concluded that, in this and similar studies, confidence in the interpretation of results is difficult to achieve. Alternative approaches may be more fruitful and merit the attention of research policy makers.