Using behavioral insights in the public sector

On April 13, 96 public servants from across the country participated in a workshop discussing behavioral insights, a growing approach in the public sector that works to inform policymakers of the human behaviors driving economic and societal outcomes.

The workshop, led by Leah Everist and Marta Garnelo Caamano, consisted of an hour-long presentation and 30-minute Q&A. Everist is a Senior Advisor at The Behavioural Insights Team (BIT), where she leads BIT's portfolio of collaborations with U.S. States. She works with leaders and teams around the country from a variety of policy areas on using behavioral insights to improve the inclusiveness and effectiveness of programs that serve the public. Garnelo Caamano is a Principal Advisor at BIT Americas. She works with government agencies and international organizations to design interventions informed by behavioral science and rigorously evaluate their impact to inform policy decisions. 

The two began the workshop by introducing what they do at BIT and highlighting how their work interacts with the public sector. Garnelo Caamano explained that BIT relies heavily on behavioral science, rigorous evaluation, and capability research.

“BIT is a social purpose company. We were born as part of government and it’s part of our DNA to be government-focused,” Garnelo Caamano said. “Our mandate is to apply behavioral science [to determine] how we can bring that knowledge to those disciplines that study human behavior to improve policy and service design.”

How do behavioral insights work?

After introducing their work, Garnelo Caamano invited attendees to participate in an exercise that demonstrates how behavioral insights operate. She read off a list of several words for participants to remember such as “bed,” “blanket,” “slumber,” “yawn,” and “tired.” She then asked how many people recalled the word “sleep.” Despite many participants raising their hands, Garnelo Caamano reminded participants that this word was not on the list, and that this false recollection is an example of pattern recognition.

“What your mind is doing there is something really smart. It is engaging in pattern recognition. All of these words have something to do with sleep, even though the specific word “sleep” is not on there,” she said. 

In another example, Garnelo Caamano shows that the choices presented to us influence our preferences. In one study, a group of MBA students was asked which kind of news subscription they would prefer, a print subscription, a cheaper online subscription, or a print and online subscription for the same price as the print subscription. While participants, a group of MBA students, chose the subscription with both print and online access, when the options changed and just print and online were available, they opted for just the online subscription. 

Garnelo Caamano then applied this work to the public sector and policymaking. She emphasized that understanding how people behave in practice helps us design better policies and services. While decision makers often opt to simply implement a program they have faith in, Garenelo Caamano notes that rigorous evaluation and testing are essential for the success of a new policy.

“The key message here is that it is important to evaluate the impacts of interventions partly because what may seem like common sense or a sound hypothesis may not turn out to be the case,” she said. “There are a lot of examples of public policies with things that had good intentions behind them and in the end, they didn’t have the impact intended.”

The EAST Method 

With this foundation established, Everist began to present the EAST method to participants. This tool was developed by the BIT to capture the fundamental elements of behavior change. When one wants to shift behavior, one should make the target behavior easier, attractive, socially compelling, and timely (EAST). 

In outlining this framework, Everist provided examples of behavioral insights at the local government level. For example, when discussing the “easier” component of the EAST method, Everist explained that it is essential at this stage to set intelligent defaults, reduce the hassle, and simplify processes and messages. One example she gave was the government’s efforts to boost flu vaccination rates among the elderly in the United Kingdom. They evaluated reducing the friction costs of having to travel to get a flu vaccine, and it found offering people the flu vaccine as part of a home health check increased uptake compared to sending out invitation letters for a flu vaccine appointment at the health clinic.

“So just removing some of those steps where they have to make an appointment, the steps where they have to get to the appointment, were really effective in getting a lot more people their flu shot,” she said. 

Everist continued with the “attractive” component of the model, explaining that you want your targeted behavior to attract attention, feel personalized, and use rewards and incentives. To exemplify this, she told the story of how Chicago’s government improved lead testing rates by making the testing kits more visually attractive and easier to use. 

Moving to the “social” part of the framework, Everist explained how to make decisions and behaviors more social. She defined the importance of social norms, or what most people think is the right thing to do in a given situation. 

“We generally don’t like to deviate too far from what others are doing, especially in a context with other people we know and respect,” she said.

Falling under this category are also networks, or what a few special people are doing, and commitments and reciprocity. Oftentimes we have the strong urge to repay someone when we receive something. People often stay true to their behaviors tied to commitments because they feel obligated to please others in their network. 

Finally, the “timely” element of the approach follows the idea that it is important to consider the most appropriate time to reach someone when you want them to engage in a particular behavior. For example, to help people follow through on a license renewal deadline, BIT partnered with the city of Denver to design planning prompts and let them know they can renew their license online. Individuals received mailer reminders six weeks before their licenses expired. Individuals received one of three mailers, one simply notifying them of their need to renew, one with a checklist outlining the steps to renew their license, and finally, a mailer with a planner where the individual has to schedule set dates and times to complete certain parts of the renewal process. This final postcard increased online renewals by almost 8% or 9000 Denver residents.

Q&A Session

Before moving to the Q&A session, Everist and Garnelo Caamano outlined three takeaways from their presentation:

• Behavioral insights can help us identify the predictable patterns in how humans interact with policies and programs - and to work with those patterns, biases, and heuristics rather than against them
• Evaluation can help us narrow in on what works, when, and for whom
• Driving rigorous impact is easier than one might think

In the Q&A session, state workers were given the opportunity to discuss behavioral insights and how it related to their own local projects. One participant asked if behavioral insight is a part of business intelligence.

“If we are thinking about how to better use data there is definitely an overlap,” Garnelo Caamano said. “They are related in the sense that oftentimes the initial exploration we look at what data we have on this problem, on this policy, how are people reacting … Using administrative data can help you target what population you should focus on or where you will find success in shifting the behavior.

In a participant survey conducted after the workshop, 100% of participants would recommend the workshop to a friend or colleague. 96% of participants said they were likely to use what they learned in their work. Here is what some of our participants had to say about the workshop:

“This lesson had a very broad insight on many different areas of research on human behavior.” -New Jersey attendee, mid-career (10 to 20 years)

“The training was very insightful, the lecture had clear objectives, and the information was easy to follow.” -Colorado attendee, early career (less than 10 years)

Watch the recording of the workshop here.