Base-rate neglect as a function of base rates in probabilistic contingency learning.

Area of Science:

• Cognitive Psychology
• Behavioral Economics
• Decision Science

Background:

• Humans often exhibit base-rate neglect, overweighting specific predictor information while underweighting general base rates.
• Previous research demonstrated base-rate neglect when predictors are physically similar to criterion events (Goodie & Fantino, 1996).

Purpose of the Study:

• To investigate base-rate neglect in human predictions when predictors share skewed base rates with criterion events.
• To extend findings on base-rate neglect to situations where similarity is defined by base-rate frequencies rather than physical resemblance.

Main Methods:

• Utilized a matching-to-sample paradigm in an operant analysis framework.
• Conducted partial replications of Goodie and Fantino (1996) studies.
• Examined participants' predictions based on probabilistic predictors associated with criterion events.

Main Results:

• Confirmed base-rate neglect, with participants overweighting predictor information.
• Demonstrated that participants' predictions were biased toward the more frequent criterion event when following a more frequent predictor, and vice versa.
• Showed that skewed base rates in predictors, similar to criterion events, did not mitigate base-rate neglect.

Conclusions:

• Findings support the concept of pseudocontingencies, where base-rate influences significantly impact contingency learning.
• Base-rate neglect persists even when predictor and criterion base rates are aligned in their skewness.
• Human predictive behavior is susceptible to biases that favor salient predictor information over base-rate probabilities.