This long-awaited paper from Kessler and Roth was just published last week. As Roth notes, they began work on the project in 2010, when Kessler was "just a kid." In fact, I saw Kessler present this paper around that time and can confirm that he (and I!) were much younger then.
Here's the abstract:
Abstract: The United States has a severe shortage of organs for transplant. Recently—inspired by research based on hypothetical choices—jurisdictions have tried to increase organ donor registrations by changing how the registration question is asked. We evaluate these changes with a novel "field-in-the-lab" experiment, in which subjects change their real organ donor status, and with new donor registration data collected from US states. A "yes/no" frame is not more effective than an "opt-in" frame, contradicting conclusions based on hypothetical choices, but other question wording can matter, and asking individuals to reconsider their donor status increases registrations.
As a bit of background, some early work using hypothetical donor registrations provided evidence that choice frame– specifically, requiring subjects to say either "yes" or no"no" to organ donation, as opposed to an opt-in framework — could dramatically increase organ donor registration rates. For example,
Johnson and Goldstein (2003; 2004) found that 42% of their experimental subjects reported a willingness to register as an organ donor when asked to opt-in to a donor registry but that the rate was a dramatically higher 79% when subjects were required to say either “yes” or “no” to organ donation, a staggering 37 percentage point treatment effect.
Roth and Kessler use a controlled experiment (giving Massachusetts residents the opportunity to change their organ donor registration status) that tests whether the choice frame affects organ donor registration. They then present results from newly collected data on registration decisions from state DMVs, leveraging state-level changes in question wording as natural experiments. "Both sets of empirical work find consistent, near zero impacts of choice frame on organ donor registration, suggesting that prior estimates using hypothetical data do not accurately describe behavior under actual incentives."
Read the whole thing here.
Papers can take a long time to publish, but the problems they address can last even longer. #econsky #transplantation marketdesigner.blogspot.com/2025/04/newo…
— Al Roth (@alroth.bsky.social) April 30, 2025 at 8:12 AM