Article

Relativity of response rate and reinforcement frequency in a multiple schedule

Authors:
To read the full-text of this research, you can request a copy directly from the author.

No full-text available

Request Full-text Paper PDF

To read the full-text of this research,
you can request a copy directly from the author.

... Esses dois tipos, por sua vez, são classificados em (1) positivo e (2) negativo. Sendo assim, quatro tipos de interações comportamentais são possíveis: (1) contraste positivoocorre quando há um aumento na taxa de respostas no componente estável em função da diminuição da taxa de respostas no componente alterado; (2) contraste negativoocorre quando há diminuição na taxa de respostas no componente estável em função do aumento na taxa de respostas no componente alterado; (3) indução positivaocorre quando há um aumento na taxa de respostas no componente estável em função do aumento na taxa no componente alterado; (2) indução negativaocorre quando há diminuição na taxa de respostas no componente estável em função da diminuição na taxa no componente alterado (Pear & Wilkie, 1971;Reynolds, 1961b). ...
... Estudos experimentais têm demonstrado a produção de interações comportamentais em contingências apetitivas, com diversos tipos de sujeitos (e.g., pombos, ratos, macacos), utilizando esquemas múltiplos (e.g., Bloomfield, 1967;Bradshaw, 1975;Caplan & Graefe, 1980;Catania, 1961;Catania & Gill, 1964;Hemmes & Eckerman, 1972;Gutman, 1977;Gutman, Sutterer, & Brush, 1975;MacDonall & Marcucella, 1978;Pear & Wilkie, 1971;Reynolds, 1961aReynolds, , 1961bReynolds, , 1961cReynolds & Limpo, 1968;Spealman, 1978;Swindell, McSweeney, & Murphy 2003;Thomas & Cameron, 1974;Topping & Larmi Jr, 1972). Interações comportamentais também foram demonstradas utilizando contingências aversivaspunição e reforçamento negativo (e.g., Azrin, 1956;Brethower & Reynolds, 1962;Coates, 1972;Crosbie, Williams, Lattal, Anderson, & Brown, 1997;Dinsmoor, 1952;Herrnstein & Brady, 1958;Honig & Slivka, 1964;Lattal, 1970;Lattal & Griffin, 1972;Rachlin, 1966;Tullis & Walters, 1968). ...
... As variáveis que contribuem para a produção de indução, contraste ou ambos quando punição é utilizada, não são claras. Uma das variáveis importantes para a produção de contraste, por exemplo, é a taxa de reforçamento (Catania, 1961;Reynolds, 1961a;1961b;Reynolds & Limpo, 1968). Contraste ocorre quando há uma relação inversa entre a taxa de respostas no componente estável e a taxa de reforçamento no componente alterado (MacSweeney & Norman, 1979). ...
Article
Full-text available
Interação comportamental é um fenômeno que pode ocorrer em esquemas múltiplos, em que a taxa de respostas em um componente pode mudar em função de alteração do esquema de reforçamento ou de punição no segundo componente. Interações comportamentais podem ocorrer quando há utilização de contingências apetitivas ou aversivas, como punição. Com relação às contingências punitivas, existem dados conflitantes na literatura sobre a produção de contraste ou indução com não humanos, utilizando majoritariamente choque elétrico como estímulo punidor. O objetivo deste estudo foi verificar se ocorreria algum tipo de interação comportamental em esquemas múltiplos, utilizando o jato de ar quente (JAQ) como estímulo punidor. Dezesseis ratos albinos foram expostos às fases: Pré-Punição, Punição e Pós-Punição. Nas três fases vigorou um esquema múltiplo com dois componentes (claro e escuro). O esquema de reforçamento em vigor em ambos os componentes foi intervalo variável 30s (VI 30) para respostas de pressão à barra, em todas as fases. Na fase punição, o JAQ foi administrado contingente a cada resposta de pressão à barra no componente claro para metade dos sujeitos, e no escuro para a outra metade. Ocorreu indução negativa no componente não punido, para todos os sujeitos, principalmente nas sessões iniciais; ao longo das sessões, ocorreu recuperação na taxa nesse componente para todos os sujeitos, resultando em ocorrência tardia de contraste positivo para três dos sujeitos. A produção de contraste parece estar relacionada ao número de sessões de exposição ao punidor.Palavras-chave: interação comportamental, punição, contraste, indução.
... A questão agora é se a dependência do contraste ao reforço está em concordância quantitativa, bem como qualitativa, com a Equação 20. Reynolds (1961c), utilizando esquemas múltiplos VI FR, concluiu não apenas que a taxa relativa de respostas seria uma função da taxa relativa de reforços, mas também que a função seria linear, com um intercepto positivo e inclinação menor que 1,0. Isto é diferente de esquemas concorrentes, onde a relação de igualação prevalece (inclinação = 1,0; intercepto = 0). ...
... Para qualquer valor de m, a função se torna mais íngreme quanto maiores os valores de R o . Enquanto m for menor que 1,0, a função se aproxima assintoticamente da linha de igualação (7) Pontos nas funções na Figura 12 podem facilmente ser tomados como lineares, particularmente se eles se encontram no meio da amplitude, como o fez Reynolds (1961c). Um conjunto de dados mais completo para testar a Equação 21 é o do experimento de Reynolds (1963b) sobre esquemas múltiplos VI VI. ...
... Nevertheless, it is interesting to note that the highest CRF rate of any S (including the straight CRF subjects) was produced when this schedule was paired with extinction. This finding would be expected from Reynolds (1961aReynolds ( , 1961b) study of the relationship of reinforcement frequency and behavioral contrast. These data and the comparisons must be considered to be tentative since, as a consequence of the experimental procedure, rate was not independent of the number of anticipations S made. ...
... As Herrnstein and Brady (1958) point out, the effects of schedules upon behavior can be studied in a short time through the use of multiple schedules. However, there appears to be an interaction between the components of a multiple schedule (see also Reynolds, 1961Reynolds, a, 1961b. This is a limiting factor in generalizing the effects of a component in a multiple schedule to its effects in isolation. ...
Article
Full-text available
The present study was concerned with the effects of schedules of reinforcement upon the rate of verbal responding to written material in children. Four multiple schedules were used; multiple CRF-EXT, multiple CRF-VR, multiple CRF-VI, and multiple VR-VI, one subject being run on each schedule. Rates under CRF were lower than under VR, and somewhat higher than under VI, and much higher than under extinction. The subject run on multiple VR-VI showed little rate difference in the two components.
... The reduction of two absolute response rates to a relative-rate measure creates special difficulties when one of them approaches zero: the relative rate of the other approaches 1.0 no matter what its absolute rate. If relative response and reinforcement measures are invoked to account for a phenomenon characterized in terms of a change in absolute response rate, the phenomenon does not even appear in the relative measures (Reynolds, 1961). For example, in the transition from multiple variable-interval variableinterval (VI VI) to multiple variable-interval extinction (VI EXT) schedules, behavioral contrast has been defined as an increase in the responding maintained in the uncharged VI component that occurs when the other component is changed from VI to EXT. ...
... Experimentos manipulando diversos tipos de esquemas em um dos componentes, para produzir diferentes taxas de respostas sem alterar as taxas de reforços entre os componentes foram realizados, corroborando a ideia de que a taxa de reforços é realmente mais relevante do que a taxa de respostas na determinação do contraste (e.g. Bloomfield, 1967;Nevin, 1968;Reynolds, 1961d;Zuriff, 1970). Dado, então, que a taxa relativa de reforços é uma variável relevante para a ocorrência do contraste, também foram estabelecidas relações entre contraste e a Lei da Igualação (e.g. ...
... Perhaps the most well-known explanation is an extension of reinforcement interaction outlined in Chapter 3. Based on experiments that assessed response rate in one schedule component when the reinforcement rate was manipulated in an adjacent component (Reynolds, 1961a(Reynolds, , 1961b(Reynolds, , 1961cBloomfield, 1967;Nevin, 1968), the reinforcement model specifically suggests that the response rate in one component depends upon the relative reinforcement rate (Herrnstein, 1970;Catania, 1973). However, the reinforcement model of multiple schedules does not easily account for response changes occurring in one component when the response rate but not the reinforcement rate changes in the adjacent component (Brownstein and Hughes, 1970; Brownstein and Newsom, ...
Chapter
Full-text available
The response rate in the presence of one stimulus and its associated schedule depends also upon the schedules associated with temporally adjacent stimuli. Alternations between schedule components associated with different stimuli are traditionally referred to as multiple schedules (Ferster and Skinner, 1957). Reynolds (1961b) suggested that interactions in multiple schedules might be described as contrast or induction, depending upon whether the changes in response rate diverge or converge across schedule components. Furthermore, contrast and induction may be either positive or negative, depending upon direction of change in response rate. Positive contrast, for example, refers to an increase in response rate in one schedule component simultaneously with a decreased rate in an adjacent component.
... If opposing rate changes in the two components of a multiple schedule accompany the direct manipulation of the rate of responding in one of the components, behavioral contrast is said to occur. Contrast is reliably produced when the rate of responding is altered by a change in the rate of reinforcement (Reynolds, 1961a, b) .. Reynolds (1961b) has proposed the relative rate of reinforcement as a primary determinant of contrast. For components of equal duration, relative rate of reinforcement in a given component is equal to the percentage of the total reinforcement provided in that component. ...
Article
After responding was maintained on multiple fixed-interval schedules of reinforcement, a cuing procedure was added to one component. Cuing was accomplished by illuminating a lamp just prior to reinforcement availability. This procedure produced positive behavioral contrast. The rate of responding in the cued component decreased, and the rate of responding in the uncued component increased. When the cue was removed, negative behavioral contrast occurred. Rate of responding in the component from which the cue had been removed increased, and rate of responding in the other component decreased. Throughout the experiment, rates of reinforcement in both components were held constant.
Article
Full-text available
Six homing pigeons were trained on a variety of concurrent-chains schedules in which the initial links were equal variable-interval schedules and the terminal links were fixed-interval schedules. Both terminal-link and initial-link schedules were systematically varied. The results showed that preference for a particular terminal-link schedule combination was greater, the shorter the initial-link schedules. The data closely matched predictions from the model of choice suggested by Davison and Temple (1973), but did not match predictions from two other models. An alternative method for analyzing concurrent-chains performance by assuming that the schedule consists of both chained schedules and successive, discriminated components that comprise multiple schedules, was suggested.
Article
When the consequences of a response become less reinforcing in the presence of one stimulus, reinforced responding in the presence of a second stimulus often increases. This phenomenon is called positive behavioral contrast. This research evaluated the effects of an analogous manipulation in an applied setting within the context of a classroom token economy. Seven severely to profoundly hearing impaired middle school students who were disruptive and low-achieving in math class served as subjects. After a pretoken economy baseline, token reinforcement was instituted at three learning stations. Reinforcement was then withdrawn from two of the three stations but continued at the remaining one. The results indicated that academic response rate increased at the station where reinforcement was continued while it decreased at the two stations where it was withdrawn. These effects were reversed when reinforcement was reinstated at all three stations. The reciprocal relationship between increases in academic performance and decreases in measures of classroom disruption previously reported in the literature was also replicated.
Article
Pigeons were trained on a multiple schedule in which the duration of access to grain reinforcement was varied independently in the two components. The relative response rate in one component was an increasing function of the relative duration of reinforcement in that component. The similarity of this interaction to that found in multiple schedules of different reinforcement frequency is discussed. Extinction data were also similar to those obtained after training on multiple schedules of different reinforcement frequency.
Article
Full-text available
Pigeons' responses on either of two concurrently available keys each associated with variable-interval 60-sec schedules occasionally changed the schedule on that key to a terminal-link interval schedule providing access to gain while the other key became inoperative. Experiment I compared simple fixed- and mixed-interval schedules in the terminal links, and showed that for all subjects and schedules the distribution of responses during the concurrent initial links was accurately described by the relative inverse delay of reinforcement squared. Experiment II extended the generality of this formulation to a situation in which rate of reinforcement was constant and delay alone was varied.
Chapter
This chapter discusses the by-products of discrimination learning. When an organism learns discrimination, the discriminative stimuli may acquire functions other than their basic function of occasioning differential responding. In addition, the discriminative stimulus correlated with extinction, or with a relatively poor schedule of reinforcement, may acquire aversive, inhibitory, and emotional functions. Discrimination learning may also result in an increase in the excitatory potential of the stimulus correlated with the richer schedule of reinforcement. The nondiscriminative functions of discriminative stimuli that are by-products of the formation of discrimination are illustrated in this chapter. These by-products of discrimination learning are neither universal nor permanent characteristics of discrimination learning.
Article
Three pigeons were trained on a multiple VI-3 FR-50 schedule of reinforcement. An Estes-Skinner conditioned suppression procedure, with a 1 min. CS, was then superimposed on the VI component. After the suppression was stable, the number of responses required per reinforcement in the FR component was progressively increased to produce an increase in the response rate during the VI component. The results indicated no change in suppression accompanied by substantial changes in the baseline rate of response during the VI component as the FR requirement was increased. These data were interpreted to indicate that an increase in the baseline response rate is not a sufficient condition to produce a change in conditioned suppression on a variable interval schedule.
Article
Four rats were trained to escape electric shock by satisfying various negative reinforcement schedules. The schedules employed included continuous escape, fixed interval (FI), differential reinforcement of low rate (DRL), differential reinforcement of other behavior (DRO), and tandem fixed interval differential reinforcement of high rate 4 escape (tandem FI DRH 4 escape). Various combinations of these escape schedules, accompanied by discrete auditory and visual cues were alternated at 10-min intervals in a multiple schedule format. Each escape schedule was linked with a 20-sec avoidance component (R-S = 20 sec), in which responses prevented the occurrence of escape periods. In general, all rats produced the high or low escape rates expected from the escape schedules employed. The data produced during the shock off intervals revealed that avoidance was higher in conjunction with the fixed interval and DRL escape schedules relative to the avoidance produced when the continuous and DRO escape schedules were employed. The differential escape rates, in conjunction with the differential avoidance performances, indicated evidence of stimulus control by the various cues associated with the various escape-avoidance components employed.
Chapter
Reinforcement schedules have been perceived as simply more complex extensions of situations that have always been used for the experimental investigation of learning. However, they seem to be distinctive in at least two ways. Reinforcement schedules are typically programed automatically, and animals are exposed to a given procedure for a protracted period. Because of the automatic scheduling of events by clocks and counters, animals exposed to reinforcement schedules have many opportunities to come under the control of temporal regularities of various sorts. Under these conditions, it is easy to become aware of the important role of temporal variables in the determination of adaptive behavior. The chapter presents the idea of reinforcement as selection of behavioral variations, within the limitations of current knowledge.
Chapter
Operants are constructed by reinforcement contingencies. The chapter describes the procedures for inducing new response topographies, including reinforcement, discriminative stimulus control, deprivation, reflex elicitation, release of instinctive response patterns, and emotional induction. Emotional induction covers operations that are insufficient by themselves to guarantee the occurrence of any particular topography but that set the stage for a certain range of stimuli to be effective in inducing a delimited range of topographies. Sometimes the effects of emotional induction take the form of modulating the direct effects of reinforcement contingencies on operant behavior, and sometimes they take the form of inducing wholly new and unexpected topographies. Examples of emotionally induced topographies include schedule-induced polydipsia and pica; shock-induced copulation, eating, and aggression; extinction-induced aggression; behavioral contrast; and distress vocalizations induced by removal of an imprinted stimulus. Not all induced topographies are equally amenable to becoming operants. The ease of rendering a topography operant may be related to the ease with which the topography can free itself of control by its inducing stimuli.
Chapter
Several important characteristics of reward training have been revealed by employing procedures in which animals are exposed to more than one value of the reward event. The present chapter reviews three of those procedures and the contrast effects they produce in instrumental performance. “Contrast effect” is a term borrowed from sensory psychology where it describes the fact that the perceived difference between two stimuli is exaggerated by the manner of their presentation. A contrast effect in reward training is characterized by performances which indicate that the influence of a given reward event is exaggerated by the nature of other reward events to which the animal is exposed. The procedures to be discussed here yield performances that have been designated successive-, simultaneous-, and behavioral-contrast effects. They have been an important stimulant for theoretical development.
Article
Three pigeons received food independently or dependenly of key pecking according to a two-component multiple schedule, with an equal frequency of food in both components. Simultaneously, during only one of the components the birds received shocks either independently or dependently of key pecking. The location of stimuli associated with each multiple-schedule component was also varied. Behavioral contrast was observed only when discriminative stimuli were on the instrumental key and shock was presented after each key peck. There was no evidence of pecking at the signal associated with the absence of shock when food and shock were presented independently of responding, or when the signal associated with the absence of shock was presented on a second key.
After responding was maintained on multiple variable-interval schedules of reinforcement, a signaling procedure was added to one component. The signaling procedure consisted of illuminating the key, the only source of illumination in the chamber, only when responding would be reinforced. Rate of responding in the unaltered component increased. When the signaling procedure was removed, rate of responding decreased in the component in which reinforcement had never been signaled. Obtained rates of reinforcement in both components were equal throughout the experiment.
Article
Three white Carneaux pigeons were trained to respond on a mult VI 1-min. (milo reinforcement), VI 1-min. (pea reinforcement) schedule when each component was associated with a different key, feeder, and reinforcer. The experiment was divided into four phases. In Phases 1 and 3, baseline rates of responding were established. In experimental Phases 2 and 4, one component of the multiple schedule was changed to extinction. During the experimental phases, response rates decreased in the extinction component and increased in the unchanged component (positive behavioral contrast). The increase in responding in the unchanged component was greater when the more valued reinforcer was extinguished. These findings are very similar to those reported by Beninger and Kendall (1975) and extend the positive contrast effect to another species, pigeons.
Article
The rate of pecking under the variable-interval component of a multiple variable-interval extinction schedule was investigated as a function of whether the components were changed after reinforcement or were changed on the basis of elapsed time. Behavioral contrast—an increase in the response rate during the unchanged variable-interval component—was observed when the components were changed on the basis of time, but not when the components were changed after reinforcement. These results support the hypothesis that behavioral contrast is primarily a result of selective reinforcement of short interresponse times.
Article
Three pigeons were exposed to a procedure in which a delay was imposed between the occurrence of a response and the presentation of reinforcement. With the delay associated with one key color held constant at 3 sec, the delay associated with a second color was increased first to 7 sec and then to 12 sec. As the delay was lengthened, long response latencies to the stimulus associated with this increase became more frequent, but modal response latency was affected little. Response latencies to the stimulus associated with the unchanged delay, on the other hand, decreased as the delay in the other component was lengthened.
Article
Two white rats were run on a MULT VI 30 VI 30 VI 30 schedule of reinforcement until they achieved a steady rate of responding from session to session and from component to component. They were then run on a MULT PUN-EXT VI 30 EXT VI 30 schedule of reinforcement. Higher rates of responding were observed in the reinforcement component following PUN-EXT than in the reinforcement component following EXT. The results were discussed in relation to present theories of behavioral contrast and in relation to the aversive qualities of the suppression stimuli and the conditioned reinforcing properties of barpressing.
Article
The elimination of punishment following each response on a fixed-ratio schedule of positive reinforcement occurred whenever pigeons faded to respond on the fixed-ratio operandum for predetermined time intervals. Both increases in the time interval required for escape from punishment and repeated exposure to each escape requirement interval resulted in reductions in the occurrence of escape behavior.
Article
Full-text available
Rats were rewarded concurrently, at equal frequencies, for pushes at the doors in front of two reinforcement magazines. The duration of the reinforcer given at one magazine was constant, while the reinforcer duration at the other magazine was changed every six sessions. For three rats the constant reinforcer was 3 sec, and for three other rats the constant reinforcer was 1 sec. For all these animals the duration of the alternative reinforcer was varied between 1 and 5 sec. Rate of response at the magazine that delivered the constant reinforcer duration was found to vary inversely with the duration of the reinforcer obtained at the alternative magazine. The matching of relative response rate to relative reinforcer duration was poor, and the results are attributed partly to the general rate-suppressing effects of long reinforcer durations.
Article
A new technique was applied to the study of operant discrimination in the Skinner box. Experimental session was divided into 20 periods, half with positive stimulus, and half with negative stimulus. Responses were reinforced aperiodically during the positive periods. 16 rats were run under these conditions for 27 days original discrimination; then 33 days on the reverse discrimination; then 5 days extinction. As the discrimination formed, the over-all rate of responding remained constant, and the responses shifted from the negative periods to the positive ones. During extinction the discrimination was not abolished.
Response latencies as a measure of the interaction of components on a multiple fixed-ratio schedule
  • Schuster
Schuster, C. Response latencies as a measure of the interaction of components on a multiple fixed-ratio schedule. J. exp. anal. Behav:, 1959, 2, 259 (Abstract)
Behavioral consequences of the removal of a discriminative stimulus. Unpublished doctoral dissertation
  • R J Herrnstein
Herrnstein, R. J. Behavioral consequences of the removal of a discriminative stimulus. Unpublished doctoral dissertation, Harvard Univer., 1955.
An analysis of interactions in a multiple schedule. Unpublished doctoral dissertation
  • G S Reynolds
Reynolds, G. S. An analysis of interactions in a multiple schedule. Unpublished doctoral dissertation, Harvard Univer., 1960.
  • Ferster