Summary: Observing how others make decisions can help people make better decisions themselves.
Source: City University London
New research from the Business School (formerly Cass) at City University London finds that watching others make decisions can be an effective way to reduce cognitive biases and improve judgment.
Led by Professor Irene Scopelliti, Professor of Marketing and Behavioural Science, the team tested a novel debiasing approach based on observational learning. The paper reports experimental evidence that observing other people’s decision-making—rather than only receiving abstract instructions or feedback—can lead to measurable improvements in participants’ subsequent decisions.
The authors conducted three experiments. In each, participants made a series of judgments, completed a training or intervention intended to reduce bias, and then repeated similar judgments so researchers could measure changes in accuracy and susceptibility to common cognitive biases.
Experiment One: Comparing observational learning with other debiasing methods
Experiment One compared an observational learning intervention to three other approaches that are commonly used to reduce bias. Participants answered questions designed to reveal susceptibility to three types of bias: anchoring, social projection, and representativeness. Before repeating the questions, participants received one of four interventions: a brief five-minute break (control), a 30-minute instructional video explaining debiasing techniques (information-based), playing a decision-making video game for 90 minutes (active practice), or anonymously observing another participant play the game (observational learning).
Experiment Two: Teaching the “averaging” rule and measuring weight on advice
The second experiment focused on advice-taking and whether observation can teach an effective decision rule: averaging one’s own estimate with another person’s estimate to improve accuracy. Participants first estimated the weight of ten objects from photographs, then saw another participant’s estimate and had the opportunity to revise their own responses. The change from the first to the revised estimate was recorded as the “weight on advice.”
Participants then received one of four interventions: an informational explanation about averaging estimates, a video showing someone revising their estimate after seeing a peer’s estimate (observational intervention), both the information and the video, or neither. After the intervention, participants again estimated the objects and revised those estimates after seeing anonymous peer estimates. Researchers compared average changes in advice-taking across conditions.
A follow-up to Experiment Two replaced human advice with algorithmically generated advice (unknown to participants) to test whether observational learning would also increase the uptake of superior algorithmic recommendations.
Key findings from the three experiments included:
- Observational learning was an effective debiasing intervention. In Experiment One, observing another person’s decision process reduced participants’ susceptibility to anchoring, social projection, and representativeness biases.
- On average, the observational intervention outperformed the instructional video and the passive control condition in improving decision quality.
- In Experiment Two, observational learning was more effective than practice alone at teaching people how to use advice and increased the degree to which participants took advice.
- Combining observational learning with information-based instruction produced stronger improvements than information alone, indicating that seeing someone apply a decision rule has unique, additive benefits for learning that rule.
Professor Scopelliti emphasized the practical value of observational training as a scalable and cost-effective debiasing tool. She noted that most traditional debiasing programs focus on teaching abstract rules or giving feedback about individuals’ own choices, whereas observational learning lets people learn from others’ successes and mistakes.
“Observational learning lets us internalize decision strategies by watching how other people act, which can improve our judgments without requiring extensive feedback or practice,” Professor Scopelliti said. “Social learning interventions are promising because they are relatively inexpensive to implement and can be scaled across many contexts—from everyday choices, like what gift to buy, to high-stakes professional decisions in business, law, and public policy.”
Haewon Yoon, Assistant Professor of Marketing at Indiana University Kelley School of Business, commented that the results show the potential for observational learning to reduce decision biases and improve decision-making efficiency. She noted that when people watch others—whether in games or real work settings—they can observe the decision process and learn which strategies lead to better outcomes.
“Our research suggests that observational learning has the potential to be used to reduce decision biases and improve decision making,” the team wrote. Observing others’ errors and successes provides concrete examples that can change how people approach uncertain choices.
Professor Carey Morewedge of Boston University’s Questrom School of Business highlighted that this work extends recent advances in debiasing research, which have typically focused on feedback about people’s own decisions. “This study shows we do not always need direct personal experience to learn to be less biased; we can benefit from watching other people,” he said.
About this research
Source: City University London
Contact: Hamish Armstrong – City University London
Image: The image is in the public domain
Original Research: Closed access.
“Decision making can be improved through observational learning” by Scopelliti et al., Organizational Behavior and Human Decision Processes.
Abstract
Decision making can be improved through observational learning
The research demonstrates that observational learning can debias judgment and decision making. Single-session observational training interventions (similar to “hot seating”) can reduce cognitive biases in laboratory settings—specifically anchoring, representativeness, and social projection—and can effectively teach decision rules that increase advice taking in a weight-on-advice task (the averaging principle). These interventions improve judgment accuracy, rule learning, and advice-taking more than practice alone. Observational interventions can be as effective as information-based programs and show additive effects when combined, especially for advice taking and for accuracy when advice is algorithmically optimized. The findings suggest that explicit information can supplement and sometimes reduce the reliance on tacit knowledge gained through observation, and that observational learning represents a distinct and valuable debiasing strategy. New scales measuring individual differences in susceptibility to anchoring, representativeness heuristics, and social projection are also reported.