Biased evaluations emerge from inferring hidden causes.

Area of Science:

• Cognitive Psychology
• Social Psychology
• Decision Making

Background:

• Human judgment often involves evaluating groups based on observed behaviors.
• The impact of rare events on overall impressions is not fully understood.
• Existing models may not fully account for the influence of infrequent observations.

Purpose of the Study:

• To investigate how rare experiences influence overall group evaluations.
• To propose and test a Bayesian inference model explaining this phenomenon.
• To understand the role of inferring hidden causes in judgment biases.

Main Methods:

• Development of a Bayesian inference model to organize environmental statistics.
• Conducting eight experiments with participants observing sequences of social and non-social behaviors.
• Analyzing how participants estimate averages after observing all data versus tracking a running summary.

Main Results:

• Estimates were biased toward sparse (rare) events when participants saw all observations.
• This bias was not present when participants tracked a summary value as observations accrued.
• The model's predictions regarding the overweighting of rare events were supported.

Conclusions:

• Biases in group evaluation can stem from inferring latent causes of observed behaviors.
• Rare events may be overweighted due to the normative inference of their underlying causes.
• Understanding causal inference is key to explaining judgment biases in social and non-social contexts.