ABSTRACTThere are four stages in the acquisition of skeletal responses when employing Pavlovian methodology: contingency detection, response selection, response acquisition, and response shaping. A two‐phase output model describes in detail the underlying response likelihood changes across the first three of the stages. Phase 1 of the model applies to the first two stages while Phase 2 applies to stage three. It was shown that Phases 1 and 2 are independent by virtue of the fact that shifts in the parameters of one are not necessarily correlated with changes in the parameters of the other and the fact that independent variable manipulations which affect one Phase do not necessarily influence the other. These results imply that there is no single, continuous underlying associative strength function across trial