Confluence effects: Not here, not nowp.

R. B. Zajonc and G. B. Markus's (see record 1976-09290-001) confluence model is both a statement of a psychological theory and a mathematical operationalization of that theory. As such, it is the first formal behavioral model of the relationship between intelligence and family configuration and applies to within-family processes. However, most tests of the model have been done on aggregate data or have lacked the necessary information to do a direct test of the model. The present study applied the confluence model to longitudinal data on 311 intact families with 906 children. Data were also tested with simple linear models, and the analyses by M. L. Berbaum and R. L. Moreland (see record 1980-32378-001) and R. C. Galbraith (see record 1982-11822-001) were replicated. Results suggest the confluence model should be rejected as a useful model for these data based on principles of parsimony and predictability. Consideration of confluence in historical perspective suggests that the underlying psychological processes are not well enough understood to expect predictability from models as complex and rigid as confluence. Strategies for future model development are proposed. (39 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)