Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
Consumers receive risk information about prescription drugs from many sources, including from largely unregulated attorney advertisements. This experiment measures subjective risk beliefs about prescription drugs and subjects’ drug-taking behavior in an incentivized setting. Results of the experiment are used to estimate a Bayesian learning model to ascertain the risk levels implied by different information sources and the informational weights subjects place on these sources. Even after people have received scientifically based drug risk information from a doctor and product label, attorney litigation advertising boosts risk beliefs about side effects by an average of 16 out of 100 cases, which is roughly a 50% increase over baseline risk beliefs. Risk beliefs have a significant effect on drug-taking behavior in an incentivized environment. Experience with taking the drug also feeds back into the learning process about risk beliefs. Our experimental results help formalize how consumers update risk beliefs and suggest ways to encourage a learning environment that avoids alarmist updating based on information sources like attorney advertising.