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.


31st Annual Convention; Chicago, IL; 2005

Event Details

Previous Page


Symposium #416
Int'l Symposium - Conjunction Fallacy and Base-Rate Error: Are Those Behaviors Controlled by Reinforcement Contingencies?
Tuesday, May 31, 2005
9:00 AM–10:20 AM
Boulevard C (2nd floor)
Area: EAB; Domain: Basic Research
Chair: Josele Abreu-Rodrigues (University of Brazil)
Discussant: Gregory J. Madden (University of Wisconsin-Eau Claire)
Abstract: The literature has systematically indicated that in situations of choice and decision-making, individuals do not always present optimal behavior. That is, their choices in such situations reveal often that the available information is not used effectively such that the probable payoff is not maximized. Questions related to decision-making have called the attention of some behavior analysts, and some of them have addressed specifically two types of nonoptimal behaviors, the conjunction fallacy and the base-rate error. The conjunction fallacy is observed when an individual reports that the conjunction of two events is more probable to occur than one of the events alone. The base-rate error consists of neglecting background information (base rates) in favor of more specific ones when judging the probability of a future event. Those nonoptimal behavior have been investigated mostly by cognitive psychologists that, although have contributed to the characterization of the phenomenon, little have said about the learning history responsible for it. This symposium, on the other hand, will focus on the conjunction fallacy and the base-rate error as learned phenomena and will point out some of their environmental controlling variables.
Verbal Relations and the Conjunction Fallacy
JOSELE ABREU-RODRIGUES (University of Brazil), Ana A. Baumann (Utah State University), Pablo C. Souza (University of Brazil)
Abstract: This presentation evaluates the effects of verbal contingencies upon the conjunction fallacy. In Experiment 1, college students were trained to discriminate stimuli correlated with different probabilities of reinforcement (high or low) by means of a matching-to-sample procedure. The participants were also asked to estimate the probability of reinforcement for each sample. The attribution of points for estimating was manipulated across groups such that estimates could be accurate, inaccurate, or assystematic. In the testing session, all participants were required to choose between a compound stimulus (high and low) and either one of the individual component stimuli (high or low). The Accurate Group chose the compound stimulus in all trials while the Inaccurate Group chose the compound only when it was pitted against the high probability stimulus. The Assystematic Group showed indifference between the two alternatives. In Experiment 2, all groups were initially trained with the conjunction rule and then exposed to the training and testing sessions as previously described. The conjunction-rule training decreased the occurrence of the conjunction fallacy for the accurate and inaccurate groups, but not for the assystematic one. These findings indicate that verbal contingencies interact with nonverbal contingencies in controlling the conjunction fallacy.
The Discriminative Function of Simple Stimuli: Does it Affect the Conjunction Fallacy?
LILIAN C. RODRIGUES (University of Brazil), Josele Abreu-Rodrigues (University of Brazil), Virginia Fava (University of Brazil), Roberta L. Leonardo (University of Brazil)
Abstract: This presentation evaluates the role of historical variables (high versus low probability of reinforcement, and reinforcement versus punishment) upon the conjunction fallacy. In Study 1, the probability of reinforcement correlated with simple stimuli was manipulated across conditions. During Training, two simple stimuli were correlated with a high probability while two other stimuli were correlated with a low probability. During Testing, college students chose between a compound stimulus (high-low) and either one of the simple component stimuli (high or low). The results showed a direct relation between preference for the compound stimulus and the relative probability of reinforcement of the high stimulus. In Study 2, during Training, two simple stimuli were correlated with reinforcement (Sr) while two other stimuli were correlated with punishment (Sp). During Testing, the participants chose between a compound and simple stimuli. There were three different compound stimuli: Sr-Sr, Sp-Sp, and Sr-Sp. With Sr-Sr, there was exclusive preference for the compound; with Sp-Sp, there was exclusive preference for the simple stimuli; and with Sr-Sp, preference for the compound was greater when the simple stimulus was correlated with punishment. These findings suggest that the conjunction fallacy is affected by the discriminative function of the simple component stimuli.
Decision Making: Base Rate, Context and Degree of Exposure
ANA A. BAUMANN (Utah State University), Amy Odum (Utah State University), Shawn R. Charlton (University of California, San Diego)
Abstract: One of the errors observed in decision-making situations is the base rate neglect, which refers to the fact that when asked to judge the probability of future events, people usually neglect contextual (molar) information and overemphasize specific (molecular) information. Prior studies tried to evaluate the variables that affect base rate neglect, usually assessing participants through questionnaires. The base rate neglect has been shown to be a robust phenomenon, being only slightly reduced and hardly eliminated by a variety of factors. The present study aimed to evaluate the effects of order of information, degree of exposure to the problems and presence of instructions on base rate neglect. Specifically, the following were manipulated: the order of base rate and the case cue presentation, how many problems the participants were exposed to, and whether an explanation about base rate was given. Preliminary studies have shown no effect of those variables. However, there is some evidence that the content of the problem (if familiar or not to the participants) may reduce base rate neglect. Further studies are being conducted to evaluate if the content, along with other variables, affects base rate neglect.



Back to Top
Modifed by Eddie Soh