Behavioral contrast as a function of component duration and baseline rate of reinforcement

Contrastis a ubiquitousconceptin experimental psychology. Most well known are contrast effects with respect to sensory systems, exemplified by color and brightness contrast. In other areas of psychology, the concept of adaptation level has wide applicability. In all of these domains, the effect of a given stimulus value seemingly involves a comparison process such that the psychological impact of a stimulus is due not solely to its absolute value, but also to its value relative to that of other stimuli presented either in the same context or in the past. Although the empirical phenomenon of contrast is pervasive, whether it is regarded as a basic psychological process or, instead, is to be explained in terms of more fundamental mechanisms varies widely across different areas of psychology. REINFORCEMENT CONTRAST:A BRIEF HISTORYThe concept of contrast has played a major role in reinforcement theory, both in generating voluminous research and in posing theoretical challenges. There are two major classes of contrast, with relatively separate experimental literatures. Successivecontrast is said to occur when the level of behavior maintained by a given reward outcome varies with the nature of the reward outcome previouslypresented. The most well known example was reported by Crespi (1942). The running speeds in a straight alley of rats maintained on a constant level of reward amount were compared with those of other rats for which the same amount of reward was presented either after experience with a smaller reward or after experience with a larger reward. Subjects shifted from a smaller amount ran faster than subjects with the constant amount throughout, which Crespi labeled an elation effect, whereas rats shifted from a larger amount ran slower than those with the constant amount throughout, which was labeled a depression effect. As a result of these and similar findings, Hullian learning theory modified the concept of incentive(the K parameter) to become a performance variable, which ultimately made it difficult to distinguish the Hullian account from other competing accounts explicitly cognitive in nature. Successive contrast effects have continued to be investigated widely even with the decline in influence of Hullian learning theory, often in the context of physiological manipulations to assess their underlying basis. Because Flaherty (1996) has provided a comprehensive review of this literature, no further account will be provided here.The second type of contrast procedure is when the behavior produced by a constant reward payoff is affected by the reward conditionsthat are alternately presented for the same response under different stimulus conditions. This procedure is sometimes labeled simultaneouscontrast, a misnomer because the different reward conditions are not simultaneously present. However, the different conditions do occur in close temporal proximity, whereas in the successive contrast procedure the different reward conditions are often experienced with significant temporal separations. Although sim...

Section: Reinforcer Habituation/satiationmentioning

confidence: 99%

Behavioral contrast redux

Williams

2002

“…Second, fluctuations in responding that occur over these long periods confound the measurement of contrast. Rates of responding may double from one baseline schedule to its recovery (e.g., McSweeney et al, 1986;Spealman & Gollub, 1974), severely limiting the accuracy of quantitative statements.…”

mentioning

confidence: 99%

“…Because the size of contrast did not change with time-out duration, a second baseline schedule (multiple VI 30-sec VI 30-sec) was also conducted. The choice of the baseline schedule may intluence the size of contrast (e.g., McSweeney et al, 1986).…”

mentioning

confidence: 99%

“…Within-session procedures have been used occasionally in the past (e.g., Coelho de Rose, 1986;Williams, 1979), but most studies of contrast use across-sessions procedures . Across-sessions procedures modify the reinforcement provided by one component across successive phases from baseline to contrast and then back to baseline (e.g., McSweeney et al, 1986). Withinsession procedures measure contrast and baseline within single sessions.…”

mentioning

confidence: 99%

See 1 more Smart Citation

Positive behavioral contrast as a function of time-out duration when pigeons peck keys on a within-session procedure

McSweeney

Melville

1991

Self Cite

Three pigeons pecked keys for food reinforcersdelivered by multiple variable interval variable interval schedules in the first part of each session (baseline) and by multiple variable interval extinction schedules in the second part of each session (contrast). The variable interval schedules delivered reinforcers after an average of 4 min or 30 sec in different conditions. The duration of a time-out between the components varied in five steps from 5 10 120 sec. Positive contrast occurred for all time-out durations in both experiments. That is, the rate of responding emitted during the constant, variable interval component was greater during the contrast than during the baseline schedules. The size of contrast did not change systematically with changes in timeout duration. These results violate most theories of contrast. They are compatible with the idea that animals integrate reinforcers over intervals longer than 2 min.Positive behavioral contrast is an increase in the rate of responding emitted during a constant component of a multiple schedule with a worsening of the conditions of reinforcement in the other component. Positive contrast is usually studied by comparing responding during a baseline schedule, in which both components provide the same conditions of reinforcement, with responding during a contrast schedule, in which one component is held constant (the constant component) and the other component is changed to extinction (the changing component).Most theories of contrast predict that contrast will decrease in size if a time-out is interposed between the components of the multiple schedule. During a time-out, the stimuli that signal the components are omitted and no reinforcers are delivered. The theories differ in exactly how they make that prediction.Rachlin (1973) During the baseline schedule, operantly conditioned responses occur during both components because responses produce reinforcers. No classically conditioned responses occur because the two components provide identical rates of reinforcement. During the contrast schedule, operantly conditioned responses occur during the constant component because responses still produce reinforcers. Classically conditioned responses also occur because the appearance of the constant component represents a transition from astimulus that signals a lower valued reinforcer (extinction in the changing component) to astimulus that signals a higher valued reinforcer (a schedule of reinforcement in the constant component). If classically conditioned responses facilitate operantly conditioned responses, then response rates will be higher during the constant component ofthe contrast schedule (operantly plus classically conditioned responses) than during the constant component of the baseline schedule (operantly conditioned responses only) and positive contrast will be observed.Interposing a time-out between components should eliminate contrast because it eliminates transitions between stimuli of different values. Eliminating value transitions removes classically con...

“…Such narrow definition seemingly contradicts the empirical data: Contrast is clearly a complex phenomenon. Contrast changes as a function of the baseline reinforcement rate McSweeney et al, 1986;Reynolds, 1963;Spealman & Gollub, 1974), the availability of an alternative response and deprivation for the reinforcer produced by that alternative response Hinson & Staddon, 1978), the relation between the discriminative stimulus and the reinforcer (Hearst & Gormley, 1976;Keller, 1974;Schwartz, 1978), the form of the operant response (McSweeney, 1978(McSweeney, , 1983, the relationship between multiple schedule components (Williams, 1979(Williams, , 1980, and the duration of the multiple schedule components (Ettinger & Staddon, 1982;Hinson, Malone, McNally, & Rowe, 1978;McSweeney, 1982;McSweeney et al, 1986;Schwartz, 1978;Shimp & Wheatly, 1971;Spealman, 1976;Todorov, 1972;Williams, 1979Williams, , 1980. These variables may also interact in unpredictable ways (e.g., Dougan et al, 1986).…”

Section: Discussionmentioning

confidence: 99%

Behavioral contrast in pigeons and rats: A comparative analysis

Dougan

Farmer-Dougan

McSweeney

1989

Self Cite

The effects of reinforcement rate on behavioral contrast were examined in pigeons and rats. Each species was exposed to a series of 12 multiple variable-interval schedules, divided into four 3-schedule series. Each series consisted of a standard contrast manipulation, and baseline schedules provided a different rate of reinforcement in each of the series. The functions relating reinforcement rate to the magnitude of contrast were different across species. Rats showed a U-shaped function, with reliable contrast occurring only at high reinforcement rates. Pigeons showed an inverted U -shaped function, with contrast occurring on all schedules except the schedule providing the lowest rate of reinforcement. Pigeons discriminated between schedule components better than rats did, although differences in discrimination were probably not responsible for the differences in contrast. The results suggest that behavioral contrast in rats may be a different phenomenon from behavioral contrast in pigeons. The results cannot be explained by current theories, which view contrast as the product of a single general process.