|
| International Paper Session - Choice |
| Sunday, May 25, 2008 |
| 9:00 AM–10:20 AM |
| Chicago & Alton |
| Area: EAB |
| Chair: John R. Smethells (Central Michigan University) |
| |
| Overmatching and Optimization in a Matching Law Experiment with Human Participants. |
| Domain: Basic Research |
| LAURILYN DIANNE JONES (The Mechner Foundation), Francis Mechner (The Mechner Foundation) |
| |
| Abstract: Ten participants completed an eight-session experiment designed to study the effect that the amount of time required for the execution of the operant would have in a matching contingency situation, and the possibility that subtracting this time variable would correct for undermatching and overmatching. The operant consisted of pressing keys on the computer keyboard. Each session consisted of three blocks, each with a different set of reinforcer probabilities: 50/50, 30/70 and 10/90. Each block ended after 300 4-cent reinforcers were obtained. Within each block, the time required per operant was changed every 50 reinforcers by changing the number of keystrokes: either 2, 8 or 14 keystrokes.
In all participants, choice during the 10/90 block optimized quickly and stayed stable, with usage of the 10% operants dropping to an average of 2% or below from the fourth session on. Choice during the 30/70 block optimized more slowly, with usage of the 30% operants reaching an average of 2 to 3% in sessions five and six. Some of the participants subsequently drifted away from optimization in the 30/70 block and exhibited other effects. Factors that may have influenced these phenomena and other characteristics of the observed drifts will be presented. |
| |
| Does a Changeover Delay Increase Sensitivity to Reinforcement Allocation? |
| Domain: Basic Research |
| JAMES S. MACDONALL (Fordham University) |
| |
| Abstract: Although the literature says that in concurrent choice procedures a changeover delay increases sensitivity to reinforcement allocation, there is little empirical evidence supporting this view. In two different experiments, one using a changeover delay (MacDonall, 2003) and one not using a changeover delay (MacDonall, 2000), there was no consistent effect of a changeover delay on sensitivity to reinforcement in different rats. The following experiments used multiple schedules with identical concurrent schedules in each component but one component also used a 5 s changeover delay. The first experiment used two pairs of stay and switch random-interval schedules for the concurrent schedule. The second experiment used the standard method of two random-interval schedules for the concurrent schedule. The run lengths and visit durations were consistently longer in the component with a changeover delay, indicating the rats’ behavior was affected by the delay. In both experiments there was no consistent difference between sensitivity to reinforcer allocation in the two components. In the present experiments, using single subject’s designs, changeover delays affected run lengths and visit durations but not sensitivity to reinforcement. |
| |
| Human Choice in a Variable Environment: Local Effects of Reinforcement. |
| Domain: Basic Research |
| CELIA LIE (Victoria University of Wellington), David N. Harper (Victoria University of Wellington), Maree J. Hunt (Victoria University of Wellington) |
| |
| Abstract: Davison and Baum (2000) developed a concurrent-schedule procedure where the relative frequency of reinforcement changed a number of times within each session, and each reinforcer ratio was separated by a 10 s black-out. They found that subjects’ behaviour adapted quickly to changes in the reinforcer ratio, with subjects showing good levels of sensitivity to the reinforcer ratios (as measured by the generalised matching law – Baum, 1974). This general procedure has been used with pigeons (e.g., Davison & Baum, 2000, 2002; Landon & Davison, 2001) and also with rats (e.g., research currently underway in our laboratory). This talk will present findings from a similar procedure with human participants where participants were presented with four different reinforcer ratios (components), each separated by a 10 s inter-component interval. Participants obtained 30 reinforcers per component and 120 reinforcers in total to exit the session. Analyses will look at how participants’ sensitivities change within each component and also across each of the four different components. |
| |
| The Impact of Self-Injurious Behaviour on Tutor’s Time Allocation: An Application of the Matching Law. |
| Domain: Basic Research |
| LEWIS A. BIZO (Southern Cross University), Louisa Ellen Salmon (Southern Cross University), Tom Randell (University of Southampton), Martin Hall (University of Southampton), Bob Remington (University of Southampton) |
| |
| Abstract: Participants were asked to teach an object matching task to each of two virtual children that simulated children that self-injured using an interactive computer simulation. In two separate experiments the relative rates of self-injury that the two virtual children engaged in was manipulated across three conditions in a repeated-measures design. The proportion of time a tutor spent teaching each child was plotted as a function of the relative rate of self-injurious behaviour. Most participants under-matched and spent roughly equal amounts of time teaching each child, a few participants matched their time allocation in almost the same proportion as the rate of self-injurious behaviour. The present study represents an application of Herrnstein’s Matching Law to an applied problem. Implications for use of this computer simulation for demonstrating to naïve tutors the impact self-injurious behaviour can have on their time allocation is discussed. |
| |
|
| |