Familiarity Bias and Belief Reversal in Relative Likelihood Judgment

Area of Science:

• Cognitive Psychology
• Decision Science
• Behavioral Economics

Background:

• Probability theory posits that beliefs about an event and its complement should be inversely related.
• Existing research indicates that individuals often violate this principle when assessing likelihoods.

Purpose of the Study:

• To investigate the 'familiarity bias' in probability assessment.
• To examine how the method of probability judgment (relative likelihood vs. direct probability) affects this bias.
• To explore the underlying cognitive mechanisms driving these judgments.

Main Methods:

• Surveys were conducted asking participants to assess the relative likelihood of familiar versus unfamiliar events and their complements.
• Further studies compared judgments when participants rated probabilities directly versus when they chose the more likely event.
• A contingent weighting model was proposed to explain the observed patterns.

Main Results:

• Participants exhibited a familiarity bias, rating familiar events and their complements as more likely than unfamiliar ones.
• This bias was less pronounced when participants judged probabilities directly rather than comparing relative likelihoods.
• A greater proportion of subjects rated familiar events as more likely compared to assigning them higher probabilities.

Conclusions:

• The familiarity bias in probability assessment is a robust phenomenon, analogous to preference reversals in decision-making.
• A contingent weighting model, where attention is skewed towards supporting evidence, explains how familiarity increases perceived likelihood for both events and their complements.
• Ease of retrieving evidence for familiar events contributes to this attentional bias and inflated likelihood judgments.