#261 2003 · Eric Johnson & Daniel Goldstein (Columbia Business School / London Business School) · Public health / behavioral science
Germany and Austria have nearly identical cultures and organ-donation attitudes, but one country's consent rate is 12% and the neighbor's is 99% — because of one checkbox
the problem
Getting people to actively opt into a beneficial program almost always undercounts genuine willingness, because most people never act on a form they support in principle
background
European countries in the early 2000s split into two organ-donation policy regimes with the same basic option — become a donor or don't — framed in opposite directions: opt-in countries required citizens to actively register as donors, while opt-out countries presumed everyone a donor unless they actively registered a refusal. Policymakers in opt-in countries generally assumed consent rates reflected genuine public attitudes toward donation, and treated low registration as evidence people didn't want to donate.
Eric Johnson and Daniel Goldstein set out to test whether that assumption held, comparing 'effective consent rate' — the share of the population counted as donors — across neighboring European countries that shared closely similar cultures, religions and general attitudes toward organ donation but differed only in which policy default their government had chosen.
what everyone would do
Run public awareness campaigns to persuade more people to actively register as donors — the standard response to low opt-in rates, which treats the gap as a problem of attitude or awareness. It fails because most non-registrants aren't actually opposed to donating; they simply never complete the form, so a campaign that shifts opinion doesn't necessarily convert that opinion into the completed administrative act, which was never the actual bottleneck.
what they saw
Johnson and Goldstein saw that the registration rate wasn't measuring genuine willingness to donate at all, it was measuring who takes the extra active step of filling out a form — a much smaller and different population than everyone who actually supports donation. Because most people never revisit a default they aren't forced to actively change, the default itself, not the underlying attitude, was doing almost all the work in determining the measured outcome.
the move
Johnson and Goldstein's 2003 Science paper found that opt-in countries clustered at low effective consent rates, typically well under 25% and in some cases in the single digits, while opt-out countries clustered above 90%, even between neighboring, culturally similar country pairs such as Germany and Austria. They followed the observational finding with a controlled online experiment that isolated the default itself as the causal variable: with identical people facing an identical choice, an opt-in default produced roughly 42% donors, a no-default forced-choice condition produced about 79%, and an opt-out default produced about 82% — confirming the default alone, not just underlying cross-country attitude differences, drove a large share of the gap.
why it works
Most people's true preference toward donation is positive, but registering requires overcoming ordinary inertia and friction, and under an opt-in system that inertia keeps the non-donor default in place for the vast majority regardless of what they actually believe. Flip the default to opt-out and the same inertia now keeps the donor status in place instead, so an otherwise identical population produces a wildly different measured rate purely because of which state requires effort to leave. The controlled experiment made this precise: identical people facing opt-in, forced-choice, or opt-out conditions moved from roughly 42% to 82% donors on the default alone, proving the huge gap between countries like Germany and Austria was administrative, not cultural.
the payoff
The paper became one of the most-cited studies in behavioral economics and choice-architecture research, and is credited with directly influencing subsequent organ-donation policy debates and default-based nudge interventions across other domains — from retirement savings auto-enrollment to data-privacy consent design. A 2012 peer-reviewed follow-up (Davidai et al.) pushed back on treating the default as the sole explanation, finding that inertia, perceived social norms, and the different psychological meaning of 'opting out of donation' versus 'opting into donation' all contribute alongside the default itself, though the researchers still identified sheer inertia as likely the primary driver.
where it breaks
The technique only helps where there's a genuine gap between people's underlying preference and their propensity to act on an opt-in form — if people are genuinely opposed to a program, a default flip just produces high nominal enrollment among people who resent or don't realize they've been enrolled, which is a real ethical concern for a decision as consequential as organ donation. It also depends on the opt-out mechanism staying genuinely easy, visible, and free of stigma or bureaucratic friction to use; if declining is nominally available but practically hard to invoke, the default isn't capturing preference anymore; it's substituting for consent the person never actually gave, which is a different and much less defensible thing.
what came after
'Do Defaults Save Lives?' is now a foundational citation in Thaler and Sunstein's 'Nudge' and across public-policy literature on choice architecture, and remains the reference study any government or organization cites when weighing whether to redesign a program's default enrollment setting rather than simply campaigning harder for people to opt in.
references
- [1]The meaning of default options for potential organ donorsPNAS / PMC (peer-reviewed), 2012pmc.ncbi.nlm.nih.gov
- [2]Default Effect: Why the Default Almost Always WinsYu-kai Chou (Gamification Analysis), 2021yukaichou.com