New research reveals that a well-known cognitive bias in how we categorize ambiguous information persists even when there are no error signals to learn from. The study, led by Lenard Dome and Andy J. Wills, proposes a new computational model, OSCAR, which offers a unified explanation for this phenomenon.
The Research
Dome and Wills, from the University of Plymouth, conducted two experiments to investigate the inverse base-rate effect—a bias where people over-associate a rare outcome with a rare feature. This effect is typically explained by prediction error, a learning mechanism where discrepancies between expected and actual outcomes drive adjustments. In their first experiment, they used an observational learning procedure where participants observed category-label co-occurrences without making predictions or receiving feedback. In the second, they removed category labels entirely, using an unsupervised procedure. Strikingly, the bias persisted in both conditions, suggesting it does not require supervised learning or error signals.
The researchers then developed OSCAR, which integrates principles from established models and operates on self-generated feedback akin to pattern completion. OSCAR successfully reproduced the individual differences seen in humans across all three procedures (supervised, observational, and unsupervised) and performed competitively against alternative models on a large preexisting dataset. Moreover, OSCAR provided an explanation for previously unexplained eye-tracking data, a feat no other model achieved.
Why It Matters
This research challenges the prevailing theory that prediction error is necessary for category learning biases. Understanding that such biases can arise without external feedback has profound implications for how we learn and make decisions in real-world situations where feedback is often absent—such as when we learn from observation or by simply exposing ourselves to information. It suggests that our brains are wired to form category associations spontaneously, which could explain why stereotypes and biased judgments form even without direct reinforcement. This knowledge empowers you to recognize that your initial impressions might be influenced by subtle, automatic processes, encouraging you to question your snap judgments.
What You Can Do
To counteract potential biases in your own thinking, actively seek out disconfirming evidence, and practice slow, deliberate reasoning when making judgments about people or situations. Consider engaging in brain training exercises that challenge your categorization skills, as these can help you become more aware of your automatic thought patterns.
Source: arXiv q-bio.NC
Curious about your own brain? Take our free adaptive IQ test or try 306 brain training levels.
