Differential probability of equivalence class formation following a one-to-many versus a many-to-one training structure

Twenty college students were randomly assigned to 2 groups. Dependent variables were presses on a touch screen, reaction time, and number of errors during training. An index of consistent responding to comparison stimuli during testing for equivalence was calculated. The subjects were also presented with a postexperimental task in which they sorted symbols used in the experiment. The main finding was that the many-to-one training structure was superior in producing equivalence outcome as compared to the one-to-many structure. There are indices that the 2 groups did not differ with respect to naming strategies. Reaction time measures differed in the 2 groups during the mix training but not during equivalence trials.