Ambiguity Leads to Context-Specificity in Predictive Learning
Grade Level at Time of Presentation
Senior
Major
Psychological Sciences
Minor
Music (other major)
Institution
Western Kentucky University
KY House District #
48
KY Senate District #
19
Faculty Advisor/ Mentor
Sharon Mutter, PhD
Department
Psychological Sciences
Abstract
Everything learned is acquired in some form of context; however, context is not always encoded. Learned information can become context-specific once one attends to context. Ambiguity has been shown to cause attention to context. In this experiment, we used Callejas-Aguillera and Rosas’ (2010) restaurant and food-illness procedure to determine whether ambiguity in learning trials leads one to attend to, and therefore encode, context. College student participants learned food-illness associations by predicting whether the food would lead to the gastric illness and then receiving feedback on their response. The food-illness associations were presented in two different restaurant contexts. To manipulate the amount of ambiguity during learning, participants received either six blocks of true discrimination (TDTD), three blocks of pseudo discrimination and three of true discrimination (PDTD), or six blocks of pseudo discrimination (PDPD). In true discrimination blocks, all food-illness associations were continuously reinforced. The associations were partially reinforced in pseudo discrimination blocks which introduces ambiguity. Thus, the TDTD condition had no ambiguity and the PDPD condition had the most ambiguity. At the end of the learning phase, participants’ predictions for target cues were tested in their learned contexts or in switched contexts. We hypothesized that 1) participants in the conditions with ambiguity (PDTD, PDPD) would show context specificity; i.e., they would predict a higher probability of gastric illness during the test for target cues presented in the same context as the learning trials than for those presented in the switched context, 2) participants in the condition without ambiguity (TDTD) should show no attention to context during learning and thus rate the target cues the same in both contexts. Our results supported the claim that ambiguity during learning leads to context-specificity. Participants showed context effects in the conditions with ambiguity and did not show a context effect in the condition without ambiguity.
Ambiguity Leads to Context-Specificity in Predictive Learning
Everything learned is acquired in some form of context; however, context is not always encoded. Learned information can become context-specific once one attends to context. Ambiguity has been shown to cause attention to context. In this experiment, we used Callejas-Aguillera and Rosas’ (2010) restaurant and food-illness procedure to determine whether ambiguity in learning trials leads one to attend to, and therefore encode, context. College student participants learned food-illness associations by predicting whether the food would lead to the gastric illness and then receiving feedback on their response. The food-illness associations were presented in two different restaurant contexts. To manipulate the amount of ambiguity during learning, participants received either six blocks of true discrimination (TDTD), three blocks of pseudo discrimination and three of true discrimination (PDTD), or six blocks of pseudo discrimination (PDPD). In true discrimination blocks, all food-illness associations were continuously reinforced. The associations were partially reinforced in pseudo discrimination blocks which introduces ambiguity. Thus, the TDTD condition had no ambiguity and the PDPD condition had the most ambiguity. At the end of the learning phase, participants’ predictions for target cues were tested in their learned contexts or in switched contexts. We hypothesized that 1) participants in the conditions with ambiguity (PDTD, PDPD) would show context specificity; i.e., they would predict a higher probability of gastric illness during the test for target cues presented in the same context as the learning trials than for those presented in the switched context, 2) participants in the condition without ambiguity (TDTD) should show no attention to context during learning and thus rate the target cues the same in both contexts. Our results supported the claim that ambiguity during learning leads to context-specificity. Participants showed context effects in the conditions with ambiguity and did not show a context effect in the condition without ambiguity.