In the face of the significant methodological difficulties outlined above, this research entertains a plausible resolution, one that hinges on a population concept of behavioral processes. In particular, we stress Hull's natural-science interpretation of human performance. Under this view, individual behavior comprises of at least two intervening processes. The first represents a process that must be isomorphic with observable empirical relationships, such as the proportional relation between the degree of interference and forgetting. Establishing these relationships has been the goal of traditional academic psychology. But similar degrees of experimentally induced interference may impede some subjects more than others. Thus, in the natural-science view, relationships depicted in the first process are permitted to vary among individuals. This is captured by a second model for the population. Facts about the second process provide a sense of comparabilty of past baselines with future results, an invaluable guide for judging the success of replication studies.
For other researchers, the function of the second model may be extended to accommodate a more evaluative point of view in that it could provide the means for validating the meaning of the first process when between-subject treatments or individual difference covariates are available. In his criticism of the aimlessness of experimental research, Underwood (1975) asserts that every law of human behavior which experimental research has established may be vacuous unless a second process can be shown to link individual performance with individual differences. Other researchers who are inclined towards a population interpretation of behavior include Bock (1989) and Rogosa and Ghandour (1991).