Validating the confluence model.

Several recent articles (Galbraith, 1982; Grotevant, Scarr, & Weinberg, 1977; Page & Grandon, 1979; Rodgers, 1981; Steelman & Mercy, 1980; Svanum & Bringle, 1980; Velandia, Grandon, & Page, 1978) claim to have tested the confluence model (Zajonc, Markus, & Markus, 1979), and many of them report having disconflrmed it. This contention is based on discrepancies between the data of these studies and what might be expected from the predictions of the confluence model. These discrepancies are invariably of two kinds. On the one hand are discrepancies between aggregate and individual data. Thus, when aggregate data are analyzed, the variables considered by the confluence model are found to explain a substantial amount of variation in intellectual performance. How