Secondary Reinforcement Effects as a Function of Reward Magnitude Training Methods

Abstract

5 groups of 10 albino rats were trained by direct goalbox placement to either a black or striped goalbox. Two groups were trained under absolute conditions with either large (A-L) or small reward (A-S), and 2 groups were discrimination trained to the goalboxes with either large (D-L) or small (D-S) reward. A fifth group received differential training with large reward to one goalbox and small reward to the other (Df). The test period, consisting of free and forced-choice T-maze trials extended over 10 days. It was found that only the A-S mean failed to exceed chance, the Df mean significantly exceeded the A-L mean but not the D-L or D-S means, and even under highly controlled discrete-trial test conditions, differential training produces a stronger reward magnitude effect than absolute training.