Differential Reward Research Articles

Three-level differential reward magnitude discrimination in rats

In two experiments conducted in a discrete-trial fixed-ratio leverpress situation, rats clearly discriminated among three stimuli associated with large reward (5 or 11 pellets), small reward (1 pellet), and nonreward. The course of acquisition and extinction was mapped for measures of start speed, ratio speed, and foodcup entrance behavior. Throughout acquisition and in the early portion of extinction there was an inverse relationship between leverpress speed and foodcup entrance responses. Resistance to extinction varied directly with amount of reward in within-S comparisons.

Differential reward positioning and children’s performance on dimension-abstracted oddity problems

This experiment tests the effects of rewarding all three positions, or only the left and right positions, in three-stimuli oddity problems. After pretraining, Ss were assigned either to a simple oddity problem or to a dimension-abstracted oddity problem, with their nonpreferred dimension being variable and irrelevant. Rewarding all three positions in an array significantly improved performance on problems containing variable irrelevant dimensions. On simple problems, no differences due to reward positioning were found.

Modification of Conceptual Style Preference by Differential Reinforcement

The study was designed to test the hypothesis that analytic and nonanalytic responding to the Conceptual Styles Test could be modified, in either direction, by application of appropriate differential reward. 60 first- and second-grade Ss, selected as high or low in analytic responding on a pretest, were given analytic, nonanalytic, or random reward training, followed by a posttest. Results indicated that training significantly altered initial response tendencies. During training, regardless of initial predisposition, Ss increased in the type of response which was rewarded. Residual effects of training, assessed by comparing posttest with pretest performance, were greater for analytic training than for nonanalytic training.

MODIFICATION OF CONCEPTUAL STYLE PREFERENCE BY DIFFERENTIAL REINFORCEMENT

The study was designed to test the hypothesis that analytic and nonanalytic responding to the Conceptual Styles Test could be modified, in either direction, by application of appropriate differential reward. 60 first- and second-grade Ss, selected as high or low in analytic responding on a pretest, were given analytic, nonanalytic, or random reward training, followed by a posttest. Results indicated that training significantly altered initial response tendencies. During training, regardless of initial predisposition, Ss increased in the type of response which was rewarded. Residual effects of training, assessed by comparing posttest with pretest performance, were greater for analytic training than for nonanalytic training.

The frustration effect as a function of training magnitude: Within- and between-Ss designs

One group of rats (Group D) received differential reward conditioning in the first alley (AI) of the double-alley apparatus prior to the administration of nonrewarded (N) test trials in both S+ and S-. Nondiscrimination groups always received the magnitude associated with S+ (Group 80 or with S- (Group 1C) for Group D prior to the administration of N trials in A1. Speeds with which Ss ran to the second alley following N in A1 were independent of training magnitudes both within Group D and between Groups 8C and 1C.

Extinction of differential reward magnitude discrimination in a discrete bar-pressing situation

Two groups of rats distinguished by the number of bar-presses required to complete a trial (FR 1 or FR 5) were given acquisition, extinction, and reacquisition in an instrumental discrimination involving differential reward magnitude (five pellets vs one pellet). The FR 5 group displayed better discrimination and extinguished faster than the FR 1 group. A “reversed magnitude-extinction effect” was found in that the usual between-groups finding of an inverse relation between resistance to extinction and acquisition reward magnitude failed to appear, ie., five-pellet start speeds of both groups remained above one-pellet speeds throughout extinction and there was a nonsignificant tendency for relative extinction rates of start speeds to be faster under one-pellet than five-pellet cues.

Differential reward conditioning in the pigeon

Pigeons trained in discrete trials to peck a red or a green key for I or S food pellets show stable negative contrast when the two colors are differentially rewarded. The contrast effect disappears when differential reward is discontinued and reappears when differential reward is resumed. Daily doses of pentobarbital sodium (5 mg/kg) do not prevent contrast. The magnitude of the effect increases within sessions and decreases between sessions, even in highly trained animals.

Competing responses in a differential magnitude of reward discrimination

The incidence of alley retracing was recorded in a differential reward magnitude discrimination. Five groups of rats received 500 mg reward to one stimulus (S1) and either 500, 250, 90, 37, or 0 mg reward to the other (S2). Following training, extinction trials were run to S1 only. The retracing patterns observed were discussed in the context of frustration theory.

Learning of two-choice, differential reward problems with informational constraints on payoff combinations

To obtain evidence concerning the dependence of choice probability upon anticipated reward, as opposed to the cumulative effects of previous rewards, college student subjects were run on two successive sets of two-choice, differential reward problems under a noncorrection procedure. Relations between the two sets were such that transfer from first to second set learning could occur only via the learning of probabilistic contingencies between reward values. Significant transfer was observed, and error frequencies on the second problem set were monotonically related to differences in anticipated reward values rather than to differences in reward magnitudes actually received for particular responses. A four-state Markov model, embodying the assumption that reward values affect performance, especially in the intermediate states, but not the rate of associative learning, accounts reasonably well for the observed learning curves and several statistics of error data for the second problem set.

An investigation of conditions determining contrast effects in differential reward conditioning.

An investigation of conditions determining contrast effects in differential reward conditioning.

Differential reward conditioning: S− contrast as a function of the magnitude of S+

Forty-two rats received 19, 10, or 1 pellets of food reward in one (S+) of two runways of different brightness. In the other runway (S−) all Ss received 1 pellet. In the initial segments of the response chain speeds in S− were an inverse function of reward magnitude in S+. Response speeds in S+ did not differ significantly.

Effect of Prior Experience with Discriminanda on Differential Conditioning

Different amounts of reversal training: None, 4 reinforced trials to the white alley (4W+) and 4 nonreinforced trials to the black (4B−), 10W+ vs 10B−, and 30W+ vs 30B−, were given prior to differential instrumental reward conditioning, B+ vs W−. Speeds of 20 male hooded rats per group were measured on the 4 trials given per day. Both average speeds and individual Ss' discrimination scores showed that as more predifferential trials were given, more trials to discrimination were required on the reversed differentiation problem. The results were discussed in terms of some basic assumptions made by frustration theories of nonreward.

Transfer of verbal discriminations based on differential reward magnitudes.

Transfer of verbal discriminations based on differential reward magnitudes.

Effects of knowledge of results and differential monetary reward on six uninterrupted hours of monitoring. SAM-TR-65-62.

Effects of knowledge of results and differential monetary reward on six uninterrupted hours of monitoring. SAM-TR-65-62.

Conditions affecting acquisition of observing responses in the absence of differential reward.

Conditions affecting acquisition of observing responses in the absence of differential reward.

Effects of Differential Reward Magnitudes on Simple Brightness Discrimination with Free and Forced Trials

Six groups of 10 hooded rats were rewarded to both of two brightness discriminanda. Differential reward was associated with one discriminanda by delivery of 2-1, 6-3, 12-6, 3-1, 6-2, or 9-3 food pellets to the respective groups for 156 trials, 2 choice and 4 forced daily. In a second experiment, 16 of 60 Ss from Exp. 1 received 84 further trials to 6 or 1 pellets. In Exp. 1, one of 6 groups (2-1) showed differential choice of the larger reward while all groups showed strong position habits. On instrumental (forced) trials all groups ran significantly faster to their larger reward. In Exp. 2, Ss previously showing no preference for their larger reward tended toward such a preference, while Ss that had heretofore shown differential choice behavior and no differential speed did develop a significant speed differential.

Response Latency of Monkeys as a Function of Reward Amount and Trials within Test Days

Recent studies by Michels (1957) and Schrier (1961a) have demonstrated a significant relationship berween amount of reward and response latency of monkeys. In both experiments, each S received the different amounts of reward, and the reward function obtained, though showing a significant downward trend, was nonlinear. A progressive increase in response latency within test days was reported by Schrier in the same scudy and also by Michels in an earlier experiment (1955), wich a 30-sec. incer-trial interval employed in both studies. The increase in response latency occurred after an initial warm-up effect (Schrier, 196la). The present article reports the results of three experiments in which latency of object displacement by monkeys was measured, and in which different Ss received different reward amounts. Differential reward amount effects can not be obtained as readily with this procedure (the nonshift procedure) as [hey can with cbe procedure in which each S repeatedly receives the differenc levels of reward (the shift procedure), at least when the dependent variable is frequency of errors on discrimination tasks (Lawson, 1957; Schrier, 1958). Differences in the inter-trial interval employed in the experiments were expected to throw some light on the narure of the previously reported response decrements within sessions. subject^.-Thirry-four adolescent rhesus monkeys were used as Ss, 10 in Exp. I, 10 in Exp. 11, and 14 in Exp. 111. Ss used in Exps. I and I1 were experimentally naive, whereas those used in Exp. 111 had had extensive experience with object-discrimination and oddity problems. Also, eight of the latter Ss had served in an experiment involving spatial separation of cue and response (Stollnitz & Schrier, in press). None of the Ss used in Exp. I11 had received any training during the six months preceding the starc of the experiment. The deprivation level for Ss in each experiment was randomly varied between 16 and 22 hr. Apparatu.--The Wisconsin General Test Apparatus (WTGA) was used throughour, the type described by Harlow (1949) in Exp. I, and a modified version wich motorized screens, described in detail by Schrier (1961b), in Exps. I1 and 111. The test tray in each experiment contained a single food well which 'This investigation was supported by a research grant (M-2818) from the Narional Institute of Mental Health, Public Health Service.

The effect of differential reward on speech patterns

Data from this experiment indicate that verbal approval of a prior speaker led to disruptive effects on certain speech patterns of a second speaker, both in a condition where the second speaker received no verbal approval and in a condition where the second speaker received verbal approval. Subjects in these two conditions presented a significantly greater number of nonfluencies than did subjects in a condition where the prior speaker had not received verbal approval. Also, the estimates of speaking time indicate that greater anxiety may have existed in these two conditions. Unlike some previous research, this experiment yielded no significant differences in speaking rate among the three conditions.

Differential reward value of incomplete and complete sexual behavior.

Differential reward value of incomplete and complete sexual behavior.