Carnegie Mellon University
December 03, 2015

Not Mere Trickery: Effects of Behavioral Nudges Persist Despite Disclosure

By Shilo Rea / 412-268-6094 /

pen and form

Nudging people toward particular decisions by presenting one option as the default can influence important life choices. However, many policymakers and some critics of behavioral interventions have raised serious ethical concerns, arguing that nudging people toward an option without their awareness is unethical, and that defaults only work because people are not aware that they are being manipulated by them. If people knew they were being nudged, the argument goes, they would resist the influence of the default or even, possibly, react by deliberately rejecting the default. 

A team led by Carnegie Mellon University behavioral economist George Loewenstein investigated whether the common assumption that defaults don’t work if people are aware of them is true. The researchers found that warning people that they were about to be nudged, or informing them after the fact and allowing them to change their decisions, did not significantly diminish the effectiveness of the default option.

A classic example of a default being used in policy involves employees having to take steps to enroll in a retirement plan. When enrollment wasn’t the default, most employees failed to enroll in tax-protected retirement accounts, or took many years to do so, despite the large financial stakes inherent in the decision. When employees were automatically enrolled in a plan but could take simple steps to opt out, research by behavioral economists showed most of them stayed in the plan.

“People think that defaults — or nudges — exploit psychological weaknesses because they are covert, or not obvious. They also think that defaults will not work if people are aware that they are being nudged,” said Loewenstein, the Herbert A. Simon University Professor of Economics and Psychology in the Dietrich College of Humanities and Social Sciences. “These findings can help to address concerns that behavioral interventions are deceitful or manipulative.”

To understand the impact of disclosing nudges, the researchers had 758 participants complete an online form about hypothetical end-of-life care choices — their preferences for medical treatment for times when they are near death or too ill to express their wishes. First, they had to choose from one of three overall goals for end of life care: to prolong life, maximize comfort, or allow medical professionals or surrogates to decide on treatment.

Next, research participants were presented with five specific medical treatments to prolong life, such as CPR, dialysis and intensive care unit admission, and for each they indicated a preference for pursuing the treatment, declining it, or leaving the decision to a surrogate or health professional. Default options were randomly assigned to either accept or reject life-prolonging treatments. Half of the participants were told about the use of defaults before they completed the forms and the other half after. Then, all participants completed steps one and two again, but without any defaults set.

The researchers found that preferences for comfort in the general directive were so fixed that they were not affected by defaults or disclosure of defaults. However, defaults for specific treatments influenced final choices irrespective of whether the intervention was disclosed ahead of time.

“Policymakers want to influence people’s behavior in a way they can defend on TV. They commonly rely on tax incentives, but those are often less effective and much costlier than changing defaults,” said David Hagmann, a graduate student in CMU’s Department of Social and Decision Sciences. “Our findings show that behavioral nudges are effective not because they are deceptive, and therefore can be employed without ethical concerns. But this is only true if the default chosen is the one that is actually best for the individual.” 

Cindy L. Bryce, associate professor of health policy and management, medicine, and clinical and translational science at the University of Pittsburgh, added, “The findings from this study are also noteworthy because many decisions are repeated over time. Transparency in the beginning allows individuals to confirm or clarify their preferences over time, rather than react to newly discovered information based on the way choice was presented. That’s true for advance directives, where people can revisit and modify their selections for medical care, and it’s also true for everyday lifestyle-related choices, like choosing to drink more water or to exercise regularly.” 

Sachin Rajpal, an undergraduate at CMU when the study was conducted, also participated in the research.