|
| Int'l Symposium - Local Effects of Reinforcers: Reinforcer Sequences, Conditional Reinforcers, and Magnitude |
| Tuesday, June 1, 2004 |
| 9:00 AM–10:20 AM |
| Commonwealth |
| Area: EAB; Domain: Applied Research |
| Chair: William M. Baum (University of California, Davis) |
| Discussant: M. Jackson Marr (Georgia Tech) |
| Abstract: . |
| |
| Choice, Reinforcement Magnitude, and Variable Reinforcing Environments |
| CARLOS F. APARICIO (University of Guadalajara, Mexico), William M. Baum (University of California, Davis) |
| Abstract: The study of choice in variable reinforcing environments has shown regularity in the local effects of individual reinforcers on response and time allocation. In this study we manipulated the magnitude reinforcement to assess its effects on choice in a variable environment with rats as subjects. Seven unsignaled components were arranged to occur for two levers, each component presenting a one of seven different reinforcer ratios, lasting until 10 reinforcers were delivered and terminating with a 1-min blackout. According to an A-B-A experimental design, we manipulated the reinforcement magnitude across phases. In phase A, independent of the variation in reinforcer ratio, the reinforcer for the left lever was four times larger than the reinforcer for the right lever. In phase B we reversed these conditions; the reinforcer for the right lever was four times larger than for the left lever. At an extended level of analysis, choice was described well by the generalized matching law, sensitivity increasing as more reinforcers were delivered in components. Local analyses showed that the most recently obtained reinforcers had substantially larger effect on choice than less recently obtained reinforcers. That larger reinforcers produced larger and longer lasting preferences is consistent with the idea that the variables controlling choice have both short- and long-term effects. |
| |
| The Effects of Runs of Reinforcers on Local Preference |
| CHRISTIAN ULRICH KRAGELOH (University of Auckland, New Zealand), Michael C. Davison (University of Auckland, New Zealand), Douglas Elliffe (University of Auckland, New Zealand) |
| Abstract: This experiment investigated the effects of differential run lengths of same-key reinforcers on local preference. Five pigeons were trained on a concurrent VI VI procedure that arranged an overall reinforcer rate of 2.2 rpm and scheduled reinforcers between the two alternatives dependently. Across experimental conditions, the average number of same-key reinforcers was varied by manipulating conditional probabilities that a reinforcer will be arranged on one alternative given that the animal had received a reinforcer immediately prior on the same alternative. To what extent can a reinforcer adopt the function of a discriminative stimulus for the likely location of the following reinforcer? |
| |
| Every Reinforcer Counts, But Do Conditional Reinforcers Count? |
| MICHAEL C. DAVISON (University of Auckland, New Zealand), William M. Baum (University of California, Davis) |
| Abstract: Using a within-sessions reinforcer-ratio variation procedure, pigeons obtained 10 food reinforcers in each of 7 components. Maintaining 10 food reinforcers per component, we added over conditions increasing numbers of magazine-light-only presentations, which we take as conditional reinforcers. Preference pulses following conditional reinforcers were consistently less extreme, and pecks per visit were fewer, than those after food reinforcers. As the proportion of magazine-light reinforcers was increased, preference pulses after food reinforcers fell significantly, but by only a vanishingly-small amount. In this procedure, magazine-light presentations act as signals for higher rates of food reinforcers. In a second procedure, we maintained an equal number of reinforcing events of both types between alternatives in all components, but varied the proportion of events that were food reinforcers between components. Under these conditions, magazine-light presentations tend to signal the lower reinforcer-rate alternative, and preference pulses following magazine-light presentations did not occur. Our tentative conclusion is that so-called conditional reinforcers are simply signals for the availability of primary reinforcers. |
|
| |