Alvin J. North

The effect of varied reinforcement training on behavior following incentive reduction

Three groups of rats were given 50 acquisition trials with large, varied, and small reward respectively. Following acquisition all Ss received 12 small reward trials. These 12 trials constituted an incentive reduction phase for the Ss trained under large and varied reward. The incentive reduction results showed a large decrement in performance for the Ss trained under large reward, while the Ss trained under varied reward showed only a very slight decrement in performance.

Irrelevant Stimuli and Degree of Learning in Discrimination Learning and Reversal

Lawrence and Mason ( 5 ) cite evidence to show that the presence of an irrelevant stimulus dimension during learning and reversal of a discrimination retards learning. On the other hand, Elam and Bitterman ( 2 ) and Jeeves and North (4) found that the learning of a discrimination was not retarded by irrelevant stimuli, while in a later study by Wortz and Bitterman ( 8 ) a retarding effect was found. Reid (7 ) , Capaldi and Stevenson ( I ) , and Pubols (6) report animal studies in which overlearning a discrimination facilitated its reversal. The purpose of the present study was to ascertain (a ) the effect of irrelevant stimuli on the acquisition and reversal of a discrimination, and ( b ) the effect of degree of learning on reversal.

The effect of number of large reward training trials on behavior following incentive reduction

Three groups of rats were given 18 large reward, 108 large reward, or 108 small reward acquisition trials, respectively. Following acquisition all Ss received 63 small reward trials. This phase constituted an incentive reduction phase for Ss receiving large reward during acquisition. The results indicated greater disruption of performance, primarily in the start measure, during the incentive reduction phase for the Ss receiving 108 than for Ss receiving 18 large reward trials during acquisition.

Acquisition and extinction with rewarded goal box placements interpolated among training trials

In a factorial design, rats were given 15, 30, or 60 training trials in a runway with 0, 15, or 30 rewarded goal placements interpolated among runs. Then 63 extinction trials were given. Terminal acquisition goal speeds increased with number of placements and also number of rewards (Ng). Resistance to extinction was found to vary inversely with number of placements, inversely with Ng, and independently of amount of training with Ng being controlled.

EXTINCTION AFTER TRAINING WITH RUNS BEING PRECEDED BY NONREWARDED PLACEMENTS IN AN INTERMITTENTLY REWARDING SITUATION

Two groups of 20 male albino rats were given either 32 rewarded placements (Group E) in a distinctive placement box (PB) or 32 non-rewarded placements (Group C). In the training phase both groups were given two placement trials per day for 20 days. One of these daily placements was non-rewarded (a “frustration treatment” for Group E Ss) and was immediately followed by a rewarded run in a runway. The other placement, not followed by a run, was rewarded for Group E Ss and non-rewarded for Group C Ss. Following a 5-trial transition phase with rewarded runs only, both groups were given 40 extinction runs (without prior placements). The acquisition and transition phase runway performances of the two groups were very similar, but Group E Ss were more resistant to extinction. The extinction findings were predicted from the frustration theory of the partial reinforcement effect.

Secondary reinforcement effects of placement procedures in a runway situation

In Experiment 1 rats were given (a) differentially, (b) nondifferentially, or (c) nonrewarded placements interpolated among nonrewarded runs. Differential placements produced significant increments affecting only start speeds. In Experiment 2 prior continuously- or partially-rewarded placements each produced significant increments in goal speeds, but not in start speeds, and did not differ in their secondary reinforcing effects.