Children's choice: Sensitivity to changes in reinforcer density

Two experiments were carried out in which children's sensitivity to changes in reinforcer density (number of reinforcers per session) was measured in a choice paradigm. In Experiment 1, 24 girls (ages 6, 9, and 12 years) performed on concurrent-chain schedules of reinforcement. The initial links were variable-interval 10-s schedules. One terminal link always gave three tokens after 30 s, but the parameters associated with the other were varied. Independent manipulations of reinforcer size (two tokens or four tokens) and prereinforcement delay (25 s or 65 s) led to equal changes in the relative density of tokens that could be earned on the schedules. Subjects at all ages were sensitive to changes in reinforcer density brought about by changes in reinforcer size, whereas only 3 12-year-olds showed sensitivity to the changes brought about by manipulation of prereinforcer delay. In Experiment 2, titration procedures were used to test the extent of this insensitivity to delay in 32 6- and 12-year-old children. In these procedures, a repeated choice of the large reinforcer increased the delay to its delivery, and a repeated choice of the small reinforcer reduced the delay to the delivery of the large reinforcer. Whereas 6-year-old boys and girls tended to maintain a strong preference for the large reinforcer, so increasing the delay to its delivery, 12-year-olds tended to distribute their responses to both alternatives, thus producing a stable level of delay to the large reinforcer. The results from the two experiments support the idea of two stages in the development of adaptive intertemporal choice.     

A quantitative analysis of chain-schedule performance

Six pigeons were trained with a chain variable-interval variable-interval schedule on the left key and with reinforcers available on the right key on a single variable-interval schedule arranged concurrently with both links of the chain. All three schedules were separately and systematically varied over a wide range of mean intervals. During these manipulations, the obtained reinforcer rates on constant arranged schedules also frequently changed systematically. Increasing reinforcer rates in Link 2 of the chain increased response rates in both links and decreased response rates in the variable-interval schedule concurrently available with Link 2. Increasing Link-1 reinforcer rates increased Link-1 response rates and decreased Link-2 response rates. Increasing reinforcer rates on the right-key schedule decreased response rates in Link 1 of the chain but did not affect the rate in Link 2. The results extend and amplify previous analyses of chain-schedule performance and help define the effects that a quantitative model must describe. However, the complexity of the results, and the fact that constant arranged reinforcer schedules did not necessarily lead to constant obtained reinforcer rates, precluded a quantitative analysis.     

Algorithmic shaping and misbehavior in the acquisition of token deposit by rats

In two experiments, rats were trained to deposit ball bearings down a hole in the floor, using an algorithmic version of shaping. The experimenter coded responses expected to be precursors of the target response, ball bearing deposit; a computer program reinforced these responses, or not, according to an algorithm that mimicked the processes thought to occur in conventional shaping. In the first experiment, 8 of 10 rats were successfully shaped; in the second, 5 of 5 were successfully shaped, and the median number of sessions required was the same as for a control group trained using conventional shaping. In both experiments, “misbehavior,” that is, excessive handling and chewing of the ball bearings, was observed, and when the algorithmic shaping procedure was used, misbehavior could be shown to occur in spite of reduced reinforcement for the responses involved.     

Visually guided catching and tracking skills in pigeons: A preliminary analysis

Research on reaching, tracking, and catching in the pigeon has been hampered by limitations of technology. A new system was developed in which the target was a small rectangle presented on a video display terminal and the pecking response was detected with touch technology. The target moved up and down vertically with sinusoidal velocity. A coincidence between the location of the pigeon's beak and the cursor produced reinforcement. The pigeon pecked ahead and behind the target, but most pecks occurred behind the target so the dominant tracking strategy was lagging. The pigeon was adept at “catching” the target at many locations throughout the trajectory. Transfer of motor learning was tested on probe trials during which the trajectory changed from vertical to horizontal. On transfer trials the pigeons' dominant pattern of pecking immediately shifted from vertical to horizontal. The motor skill displayed by the pigeons was flexible and adaptive, suggesting that the pigeons had learned to track the cursor.     

Connectionist models of conditioning: A tutorial

Models containing networks of neuron-like units have become increasingly prominent in the study of both cognitive psychology and artificial intelligence. This article describes the basic features of connectionist models and provides an illustrative application to compound-stimulus effects in respondent conditioning. Connectionist models designed specifically for operant conditioning are not yet widely available, but some current learning algorithms for machine learning indicate that such models are feasible. Conversely, designers for machine learning appear to have recognized the value of behavioral principles in producing adaptive behavior in their creations.     

THE USE OF BEHAVIORAL ASSESSMENT TO PRESCRIBE AND EVALUATE TREATMENTS FOR SEVERELY HANDICAPPED CHILDREN

Behavioral assessment procedures were used to prescribe and evaluate treatments of maladaptive behavior for 2 children with severe multiple handicaps. In Experiment 1, the results of an assessment of reinforcer preference were used in conjunction with a functional analysis of the conditions maintaining self-injurious behavior to prescribe a treatment for a child with severe disabilities. The treatment procedure involved the use of a pressure-sensitive microswitch to activate reinforcing stimuli during two solitary conditions, during which self-injurious behavior had occurred at high rates. The results were evaluated with a multiple baseline across settings design and indicated that self-injury decreased with concomitant increases in microswitch activation. Results were maintained at 6 weeks, 8 weeks, and 6 months. In Experiment 2, the results of behavioral assessments of reinforcer preference and self-injurious behavior were combined to develop a treatment for a second severely handicapped child, who exhibited high rates of self-injury in demand situations. This treatment was evaluated with a multiple baseline across tasks design and resulted in the elimination of self-injury for up to 15 months.     

Complex decision making: A longitudinal study of process and performance

Two studies using a computer-simulated, strategy-formulation game and business students were conducted using simultaneous verbal protocols. It was found that a number of information-processing and information-evaluation thought processes were significantly related to game performance. Consistent with what is known about individual decision making, the present results suggest that, for the task used in this study, individuals who engage in causal analysis perform better than those who do not. But those who focus on negative emotions, blindly repeat previously successful decisions, and engage in illogical through processes perform more poorly than those who do not. Causal replication using real managers within organizational settings, however, is needed before firm conclusions can be drawn from this research.     

Conceptualizations of valence and instrumentality: A fourfold model

Within the framework of valence-instrumentality (VI) theory, a fourfold model is developed to predict the contribution which is made by one outcome of an alternative to the valence of that alternative. The proposed model uses four subjective conditional probabilities which are based on the presence and absence of an alternative, and on the occurrence and nonoccurrence of an outcome. Valence is defined as the difference between the attractiveness values of the occurrence of an alternative (or an outcome) and the nonoccurrence of an alternative (or an outcome). Also, differences between subjective probabilities replace instrumentalities. Alternative valence-instrumentality formulations are analyzed in relation to the proposed fourfold model. The proposed model can be understood as a consistent application of Vroom's model (1964, Work and motivation, New York: Wiley), while other models are judged to be relatively deficient. The fourfold model, in combination with a matrix formulation of expectancy concepts, addresses a number of criticisms of VI theory and makes possible a test of the theory within the limits of its own assumptions.     

THE RELATIVE MOTIVATIONAL PROPERTIES OF SENSORY AND EDIBLE REINFORCERS IN TEACHING AUTISTIC CHILDREN

We compared the effects of sensory and edible reinforcers on resistance to satiation in three autistic children while learning visual discrimination tasks. Within-subject designs were used to compare a single sensory reinforcer with a single edible reinforcer and to compare multiple sensory reinforcers with multiple edibles. Results indicated that multiple sensory reinforcers maintained responding over more trials than did multiple edible reinforcers; however, the use of single sensory reinforcers and single edibles resulted in about equal numbers of trials to satiation. Both multiple and single sensory reinforcers produced higher percentages of correct responses than edible reinforcers. The findings are discussed in terms of the advantages of sensory reinforcers in teaching autistic children.