The paradox of preference for unreliable reinforcement: The role of context and conditioned reinforcement

We discuss Belke and Spetch's (1994) work on choice between reliable and unreliable reinforcement. The studies by Belke and Spetch extend a line of basic research demonstrating that under certain experimental conditions in a concurrent chains procedure, pigeons prefer an alternative that produces unreliable reinforcement. The authors describe the variables that influence preference for unreliable reinforcement, including the signaling and the duration of the reinforcement schedules, the context in which the signaling stimuli occur, and the effects of conditioned reinforcement. Hypothetical applied examples that address these variables are provided, and their influence on preference for unreliable reinforcement in humans is discussed. We conclude by suggesting a line of applied research to examine the relationship between these variables and a preference for unreliable reinforcement.     

The role of adventitious reinforcement during differential reinforcement of other behavior: A systematic replication

Differential reinforcement of other behavior (DRO) is commonly used to decrease problem behavior by presenting reinforcers contingent upon the absence of a target response. Although it is well demonstrated that DROs decrease response rates, the processes producing these decreases are not well understood. The present study systematically replicated previous research assessing whether adventitious reinforcement of alternative behavior contributes to the effectiveness of DRO. We presented university students with two options on a computer and reinforced target responding on a variable-ratio schedule. Next, we compared decreases in target-response rates and any increases in alternative responding during DRO schedules versus yoked variable-time schedules or extinction probes. DRO schedules resulted in the lowest target-response rate and highest alternative-response rate. These findings generally provide some support for the adventitious reinforcement of “other” behavior.     

Imitation in Japanese quail: The role of reinforcement of demonstrator responding

Imitative learning has been difficult to demonstrate in animals, partly because techniques have not adequately ruled out alternative accounts based on motivational and perceptual mechanisms (Zentall, 1996). Recently, it has been proposed that differences in the effects of observation of two very different responsetopographies would rule out such artifactual, alternative accounts (Akins & Zentall, 1996). In the present research, we confirmed that strong evidence for imitation can be found in Japanese quail, and that such imitation requires the imitator’s observation ofreinforced responding by the demonstrator. Thus, under the present conditions, it appears that an observer will imitate a demonstrated responseonly if it also observes the appetitive consequences of that response.     

The role of negative reinforcement in infant caregiving: an experimental simulation

We observed 11 undergraduates in an experiment designed to simulate infant caregiving. In negative reinforcement conditions experienced by all participants, a targeted caregiving response (e.g., rocking a baby doll) produced escape from, and avoidance of, a recorded infant cry. Nine participants' caregiving was shown to be controlled by this negative reinforcement contingency. Nine participants experienced an extinction condition that consisted of an inescapable cry, and the previously reinforced caregiving responses of 2 of these participants were resistant to extinction. For both of these participants, the previously reinforced response was eliminated when an alternative form of caregiving was reinforced. These results highlight the role of negative reinforcement in infant caregiving and suggest the need for additional research on the effects of crying on caregivers as well as the development of effective strategies for minimizing infant crying.     

"Automaintenance": the role of reinforcement

A key was illuminated on the average of every 30 sec for a duration of 6 sec and this was followed by food presentations. When key pecks in the presence of the light produced immediate access to grain (autoshaping procedure) pigeons were likely to peck. When pecks terminated the keylight but prevented access to grain (automaintenance procedure) pigeons were much less likely to peck. Seven of 12 pigeons failed to develop responding during the automaintenance procedure. Four of the five pigeons that responded during the automaintenance procedure were exposed to a procedure in which responses could not immediately terminate the light. Three of the four ceased to respond during optimal automaintenance conditions, suggesting that the response-dependent offset of the keylight had been reinforcing their pecking. Responding during the automaintenance procedure was eliminated for a fifth pigeon by eliminating the contiguity of light-offset and food-onset on those trials in which the pigeon did not peck. These results suggest that: (1) automaintenance (unlike autoshaping) is not an effective procedure for reliably generating responding; (2) responding that does occur during the automaintenance procedure is reinforced by the response-dependent offset of the keylight.     

The role of reinforcement in controlling sequential IRT dependencies

Sequential dependencies were investigated with two rats in a mixed and in a tandem differential-reinforcement-of-low-rate-responding schedule. In each schedule, 5-sec and 15-sec components were presented in fixed alternation. In the mixed schedule, a 5-sec interresponse time followed a 15-sec interresponse time and a 15-sec interresponse time followed a 5-sec interresponse time in predictable sequence. The correlation between prior and subsequent interresponse times, however, existed only when the prior interresponse time resulted in reinforcement. In the tandem schedule, an interresponse time greater than 5 sec in the differential-reinforcement-of-low-rate 5-sec component was not associated directly with reinforcement. One subject demonstrated sequential response patterns similar to those noted in the mixed schedule, even though the prior 5-sec interresponse time was not reinforced in the tandem schedule. The results indicate that the prior interresponse time length alone is not sufficient to influence the subsequent interresponse time length. Implications are, however, that a temporal response pattern arises when an interresponse interacts with schedule contingencies to control the interreinforcement interval.     

The role of attentional resources in automatic detection

The question of whether an automatic detection task requires the use of limited-capacity resources was investigated in a series of dual-task experiments. The automatic task required subjects to detect a consistently mapped target digit in a display of letters. This task was paired with a variety of concurrent visual discriminations that could either occur in close proximity to the automatic target or in a disparate display area. The main finding was that accuracy in each task was higher in conditions allowing attention to be shared than when it had to be divided between separate areas. These results indicate that detection of automatic targets depends on the allocation of spatial attention to the target's area.     

Self-punitive behavior with changing percentages of reinforcement: The proper role of discrimination

Past research on self-punitive behavior has been interpreted as supporting a theoretical éxplanation based on a vicious circle rather than a discrimination hypothesis. Using rats, Melvin (1964) found that self-punitive behavior is not reduced when discrimination is aided by changing the percentage of shock trials from acquisition to extinction. This past research is inconclusive because it is based on a misinterpretation of what is the critical discrimination for extinction, namely the new response-punishment contingency. Using humans, the present study provides evidence that subjects can and do discriminate the change in percentage, but continue to show self-punitive behavior until they make the discrimination that responding results in shock rather than escape.     

The evaluation and treatment of aggression maintained by attention and automatic reinforcement.

In the current investigation, we used direct and indirect methods to assess and treat several topographies of aggression that were hypothesized to have separate operant functions in a young boy with severe mental retardation and pervasive developmental disorder. First, a functional analysis of aggression, using the methods described by Iwata, Dorsey, Slifer, Bauman, and Richman (1982/1994), was conducted and produced inconclusive results. Next, indirect methods were used to develop a second functional analysis, which showed that chin grinding (firmly pressing and grinding his chin against the skin and bones of others) persisted independent of social contingencies and that the other topographies of aggression (e.g., hitting, kicking) were maintained by social positive reinforcement (attention). A treatment designed to decrease aggression maintained by attention—functional communication training with extinction—reduced all forms of aggression except chin grinding. This latter topography of aggression, which we hypothesized was maintained by automatic reinforcement, was reduced when the response-reinforcer relation was interrupted through response blocking and the child was provided with an alternative form of chin stimulation.     

The role of parity, physical size, and magnitude in numerical cognition: The SNARC effect revisited

People indicate the physical size or the parity status of small numbers faster by a left-hand key and those of larger numbers by a right-hand key. Because magnitude information is not required for successful performance in these tasks, the presence of a number-space association (the SNARC effect) has been taken to indicate the automatic activation of numerical magnitude in all tasks with numerals. In order to test this account, in a series of five experiments, we derived two consensual markers of automatic activation of irrelevant numerical magnitude, the size congruity effect (for judgments of physical size), and the Garner effect (for judgments of parity). Both markers were found independent of the SNARC effect. Consequently, we question the traditional explanation of the SNARC effect and offer an alternative account in terms of a highly overlearned stimulus-response loop.     

Partial reinforcement in serial autoshaping: The role of attentional and associative factors

In Experiment 1, pigeons were autoshaped to two stimuli (A and B) each followed by delayed reinforcement on 50% of trials. A different stimulus (X) was presented in the interval between stimulus A and reinforcement (A-X-US) and not when stimulus A was nonreinforced (A---). The X stimulus was also presented following the stimulus B trials that were nonreinforced (B-X) and was absent when stimulus B was reinforced (B---US). Not only was the response rate to A higher, but so was the rate in the intervals immediately following A, whether or not these intervals contained X. This latter finding is inconsistent with the interpretation that pigeons discriminated between X when it followed A and when it followed B and leaves open the possibility that X acts as a “catalyst”. The same general design was employed in Experiment 2 as was used in Experiment 1 with the exception that the catalytic function of X was equated for A and B by presenting reinforcement on B-X trials. The A stimulus, however, still generated the higher response rate. A new interpretation is offered that suggests that variable reward magnitude will maintain attention to a stimulus and that attention will generate keypecking.     

Preference for unreliable reinforcement in children with mental retardation: the role of conditioned reinforcement

We examined the effects of conditioned reinforcement on children's choice between reliable (100%) and unreliable (50%) reinforcement under various stimulus conditions in a concurrent-chains procedure. The study was conducted across three experiments. Experiments 1 and 2 were conducted under conditions similar to basic laboratory work and consisted of participants selecting from one of two black boxes (placed on a table) that were correlated with different reinforcement schedules. In Experiment 3, we assessed a participant's preference for unreliable reinforcement during conditions in which the target responses were aggression and mands. Results of the three experiments showed that the participants preferred unreliable reinforcement under certain conditions. Findings are discussed regarding the role of specific stimuli (i.e., items correlated with a reinforcement schedule, adult reactions) as conditioned reinforcers and how they may influence children's preference for a response (e.g., aggression, self-injury) that produces reinforcement on a leaner schedule than a socially desirable response (e.g., mands).     

The role of locus of control of reinforcement in interpersonal attraction

The purpose of the present study was to examine initial interpersonal attraction as a function of locus of control orientation. Subjects, previously tested for internality and externality, listened to a tape of an alleged stranger of the same age and sex as the subject. In reality the tapes were prepared to be either primarily internal or external in emphasis. Subjects were then asked to make some judgments about the supposed stranger. Among these judgments was the degree of attraction felt toward the individual on the tape. It was hypothesized that similarity of locus of control orientation between a subject and a stranger would lead to greater initial attraction than would a complementary locus of control orientation. Contrary to prediction, the results showed that both internal and external subjects were significantly more attracted to an internal stranger than to an external stranger. These results were discussed within an interpersonal attraction framework which suggested the need for a two-factor theory as the basis for interpersonal attraction.     

The role of reinforcement contingencies in chronic alcoholism: An experimental analysis of one case

investigators have questioned chronic alcoholics and round that some of them have returned to normal social drinking (Davies, 1962; Lemere, 1953; Nørvig and Nielsen, 1956; Selzer and Holloway. 1957; and Shea, 1954). These findings would seem inconsistent with such notions as loss of control or craving as the basis of compulsive drinking. Moreover, antecedents and consequences of these explanatory constructs, have not been elucidated. The purpose of this research was to make an intensive study of a single subject to see if specific conditions could be identified under which a chronic alcoholic's drinking could be controlled, in a setting in which environmental variables could be systematically varied and drinking behavior documented.     

The role of automatic obesity stereotypes in real hiring discrimination

This study examined whether automatic stereotypes captured by the implicit association test (IAT) can predict real hiring discrimination against the obese. In an unobtrusive field experiment, job applications were sent to a large number of real job vacancies. The applications were matched on credentials but differed with respect to the applicant's weight. Discriminatory behavior was quantified by the extent to which the hiring managers invited normal-weight versus obese applicants to a job interview. Several months after the behavioral data were obtained, the hiring managers completed an obesity IAT and explicit hiring preference measures. Only the IAT scores reliably predicted interview decisions. More specifically, hiring managers holding more negative automatic stereotypes about the obese were less likely to invite an obese applicant for an interview. The present research is the first to show that automatic bias predicts labor market discrimination against obese individuals. Practical implications are discussed.     

The generalization of deliberative and automatic behavior: The role of procedural knowledge and affective reactions

Individuals’ goal-directed activity in one situation can affect their behavior in a later, ostensibly unrelated situation. This effect can depend on whether the latter behavior is also goal-directed or is performed automatically. Three studies showed that participants’ rate of speaking in the course of performing a speech shadowing task has a positive influence on the speed with which they later complete an ostensibly unrelated marketing survey. Furthermore, when participants’ attention is called to the goal of working rapidly on the survey, the effect of their past behavior depends on their perception of the goal's desirability. When the time to complete the survey is not mentioned, however, participants’ speed of working on the survey is influenced by their speed of speaking in the earlier situation without their awareness and it does not depend on the desirability of the goal with which their behavior is associated. An additional experiment showed that the relative impact of goal-directed and automatic processes also depends on the cognitive resources that people have available to construe the evaluative implications of their decisions.