Association for Behavior Analysis International

The Association for Behavior Analysis International® (ABAI) is a nonprofit membership organization with the mission to contribute to the well-being of society by developing, enhancing, and supporting the growth and vitality of the science of behavior analysis through research, education, and practice.


34th Annual Convention; Chicago, IL; 2008

Event Details

Previous Page


Paper Session #243
International Paper Session - Contemporary Research in EAB
Sunday, May 25, 2008
2:30 PM–3:50 PM
Chicago & Alton
Area: EAB
Chair: Jay Moore (University of Wisconsin, Milwaukee)
Structural Variables in Concurrent-Chains Choice.
Domain: Basic Research
JAY MOORE (University of Wisconsin, Milwaukee)
Abstract: Structural variables in concurrent-chains choice include the nature and loci of stimuli in the initial and terminal links. Previous research has manipulated structural variables associated with the terminal links. This research shows that preference for the more immediate of two terminal-link reinforcers is greater when the terminal-link keylights are differential rather than non-differential, and when the loci of terminal-link keylights are correlated rather than uncorrelated with reinforcer immediacy. This research will be reviewed and new data from my lab will be presented. The new research manipulated structural variables associated with the initial links in two ways. The data continue to show greater preference when differential rather than nondifferential terminal-link keylights are employed. Taken together, the data from initial- and terminal-link structural manipulations emphasize the theoretical importance of conditioned reinforcement in concurrent-chains choice. The data also suggest that subjects do not automatically prefer the more immediate terminal-link reinforcer, which quantitative models of concurrent-chains choice will need to take into account.
The Effects of Initial-Link Duration on Response Allocation in Concurrent Chains.
Domain: Basic Research
DARREN R. CHRISTENSEN (University of Canterbury), Randolph C. Grace (University of Canterbury)
Abstract: The ‘initial-link effect’ is the decrease in choice in concurrent chains for the preferred terminal link when initial-link duration increases (Fantino, 1969). We describe three experiments in which initial-link duration changed each session to test predictions of a decision model for choice (Christensen & Grace, under review; Grace & McLean, 2006). In two experiments, where initial-link duration followed an ascending and descending series across sessions, we report a bitonic rather than a monotonic trend. In the third experiment, where initial-link duration was either short or long, the decision model was successfully fit to the data. The decision model predicts the bitonic function because response strength relative to the criterion changes at different rates for each alternative. This is a novel prediction by the decision model: existing models for steady-state choice such as CCM (Grace, 1994), HVA (Mazur, 2001), and DRT (Fantino, 1969) predict that choice is a monotonic decreasing function of initial-link duration. Therefore, the decision model appears to show significant promise as a model for response allocation in concurrent chains.
Numerical Discrimination and Bisection.
Domain: Basic Research
LAVINIA CM TAN (University of Canterbury), Randolph C. Grace (University of Canterbury)
Abstract: The ability to discriminate relative numerosity and the representation of number was investigated in a relative numerical discrimination bisection task, adapted from the numerical reproduction procedure (Tan et al, 2007). Four pigeons were presented with either two or six response-dependent key-light flashes in a sample phase and, in a subsequent discrimination phase, were then required to make a choice response to indicate which number was just observed. A peck to the white key was reinforced following the presentation of two flashes, and a peck to the blue key was reinforced following six flashes. Incorrect choices were followed by correction trials. Following baseline training with two and six flashes, subjects were then tested with novel values ranging from 1-7 to examine numerical scale bisection. The nature of performance in the bisection task and its implications for numerical representation are discussed.
Errorless Learning of a Double Temporal Bisection Task.
Domain: Basic Research
JOANA ARANTES (University of Canterbury), Randolph C. Grace (University of Canterbury), Armando Machado (University of Minho)
Abstract: Our goal was to compare errorless learning and trial-and-error learning in a double temporal bisection task. During training, pigeons learned two discriminations: on Type 1 trials, they learned to choose a red key after a 1-s signal and a green key after a 4-s signal; on Type 2 trials, they learned to choose a blue key after a 4-s signal and a yellow key after a 16-s signal. Whereas one group learned these discriminations through an errorless procedure, the other group learned them through trial and error. Then, in order to analyze if the content of learning was different between groups, the pigeons were exposed to two types of tests. First, they were presented with intermediate durations, 1 s to 4 s and 4 s to 16 s, and given a choice between the training stimuli (stimulus generalization test). Next, they were exposed to signals ranging from 1 s to 16 s and given a choice between novel key combinations, for example, green vs. blue (stimulus-response generalization test). Results have important implications for two models of timing, Scalar Expectancy Theory (SET) and Learning to Time (LeT).



Back to Top
Modifed by Eddie Soh