What Do Engineers and Behavioral Scientists Have to Learn From Each Other?

Emily Roebling was a daily presence on-site as the Brooklyn Bridge grew from an idea into the eighth wonder of the world. At first, she relayed her bedridden husband’s messages. He was, after all, the chief engineer. But as the 14-year project progressed, Emily would tackle more and more of the issues herself. If it weren’t for Emily Roebling, the Brooklyn Bridge would not be standing today.  

Engineers solve problems with science. The Roeblings used physics to get the Bridge’s supports into place. Wooden caissons were floated into the East River, like boats. At the time, the caissons were heavier than anything ever floated from a U.S. shipyard. The caissons were then weighed down with masonry and sunk to the bottom of the river to create airtight chambers. Water was pumped out of these massive support structures, which were then pressurized. Workers went inside to dig, sinking the caissons below the soft river sediment and into the stable ground below. One of the caissons finally came to rest 70 feet below the surface of the East River. As a result, workers going up and down were getting decompression sickness, or the bends. The Roeblings used biology trying to limit its effects. And the Roeblings used chemistry. They chose materials that would last in the river water. In each case, the science was established. But the Brooklyn Bridge required new engineering.

It appears that Emily also used some behavioral science to get the job done. She used negotiation to try to get the best value on contracts for materials. She used power dynamics to stop the mayor of Brooklyn from removing her incapacitated husband from the project. The Roeblings did not use reference class forecasting, an idea with credible behavioral science roots. This type of forecasting neutralizes the planning fallacy, our tendency to be too optimistic with schedules and budgets. And as Bent Flyvbjerg, professor of management at the University of Oxford Saïd Business School, has shown, the benefits of reference class forecasting apply to time and cost expectations for big engineering projects. The Roeblings couldn’t use Flyvbjerg’s research, because it would come more than a century later. But today we can. And we should. Like the Brooklyn Bridge, most of our major engineering projects still take longer and cost more than promised.

Engineers who create services need to approach human behavior with the same respect with which we come to physics or calculus.

Like Flyvbjerg’s work, Shahzeen Attari’s research also shows us how valuable applying behavioral science in an engineering context can be. A professor of public and environmental affairs at Indiana University, her research gives designers a more accurate view of how the public perceives energy and water systems. And my colleagues Tripp Shealy, Elke Weber, Eric Johnson, and I are using nudges to debias engineers who want to create more sustainable infrastructure.

Another lesson from the Brooklyn Bridge is that we need to support more Emily Roeblings. She was relentlessly smart. And she refused to be bound by a single academic discipline. Roebling exemplified the best of engineering and behavioral science, though she had a degree in neither. Reflecting on her life, Roebling wrote to her son: “I have more brains, common sense and know-how generally than have any two engineers, civil or uncivil.” The part of me trained as a civil engineer agrees. Engineers are comfortable solving problems using math and science. But we think and solve more than we socialize. And working with others is how the Emily Roeblings of the world learn what she called “common sense and know-how.” Behavioral science teaches us about ourselves and others. It can help engineers learn a bit more common sense and know-how, which we can infuse into our work and teaching, without straying too far from our comfort zones (though I think we should).

The Brooklyn Bridge, DJ Neri

Behavioral science training is a necessary adaptation in the evolution of engineering. It’s the reason we started a behavioral science and engineering Ph.D. program within the Convergent Behavioral Science Initiative at the University of Virginia. But training also needs to happen at the undergraduate level and even before. There’s no excuse because engineers are always designing for people. Some undergraduate engineering programs do allow students a psychology class or two. But the goal should be to offer engineering degrees infused with entire sequences of behavioral science courses, as they already do with physics, chemistry, and biology.

Let’s not expect the same people whose goal is to generate new behavioral science knowledge to also tell us everything we can create with that knowledge.

Such a shift could unlock problems that appear intractable viewed through the lens of traditional engineering. Ogilvy’s Rory Sutherland has a great bit on how we might make trains more popular. One option is to upgrade engines and tracks so that the trip is faster. Another option is to make the ride more fun and productive. Instead of new tracks and engines, invest in plush seats, high-speed internet, even cocktails. They come at a fraction of the cost, but, with those perks, most people wouldn’t mind if the trip remained a few minutes longer. Sutherland’s playful example offers a response to what I think is the most urgent challenge for modern engineers. Available resources limit our ability to create the things we think we need, the faster trains. But these same limits often do not apply to the services we really want, better train rides. Engineers who create services need to approach human behavior with the same respect with which we come to physics or calculus.

An added benefit of bringing these two fields together is that engineers will learn that we are not immune to quirks in our own thinking and behavior. Much like Econs were the model prior to the emergence of behavioral economics, rational models of human behavior still dominate engineering design. But this is starting to change. In our research, we’ve shown that engineers succumb to endowment effects, setting more ambitious sustainability goals when they begin with points, rather than starting from zero and accumulating points for sustainability as they go. Erin MacDonald at Stanford University and Shanna Daly at the University of Michigan are doing some of my favorite work in this area, using behavioral science to help designers not only avoid biases but to find new possibilities.

It’s intuitive and useful to ask and answer, “What can behavioral science fix?” It’s much harder to ask, “What does behavioral science make possible?”

The Brooklyn Bridge also has lessons for those who identify as behavioral scientists. The Bridge is a physical reminder of the work required to go from scientific insight to practical use. Sure, theory from physics and math informed the Bridge’s record-breaking span. But it took the Roeblings’ skill and persistence to translate that theory into a bridge that was 50 percent longer than anything built before. Engineering exists because applying science requires its own concerted effort. There’s no reason scientists can’t also engineer. But let’s not expect the same people whose goal is to generate new behavioral science knowledge to also tell us everything we can create with that knowledge.

The ongoing work that gives me the most hope is not from engineering or behavioral science but from those who, like Emily Roebling, are daring enough to trespass in each area. Daniel Kahneman puts it well: “There is a technology emerging from behavioral (science). It’s not only an abstract thing. You can do things with it.” I agree. When you nudge (and think and steer), you also engineer. In my view, behavioral scientists at ideas42 (Disclosure: ideas42 is a partner of the Behavioral Scientist), Behaviour Innovation, and the World Bank are using a problem-solving process akin to the one the Roeblings used to span the East River. Those applying behavioral science today can learn from the long history of problem solving in engineering.

Let me close with a question my advisees have tired of hearing—especially since I’m no better at answering it than they are. How might we move from solving problems with behavioral science to finding new ways to “do” with this science? It’s intuitive and useful to ask and answer, “What can behavioral science fix?” It’s much harder to ask, “What does behavioral science make possible?” It’s a question that requires as much persistence as insight. But we’ve seen what happens when we ask these questions of the physical sciences: Manhattan gets connected to Brooklyn, way back in 1883, and with an iconic bridge that safely moves people to this day. What feats of behavioral science and engineering from today will still be moving people a century from now? What will it mean to move people? These are questions that engineers and behavioral scientists can only answer by working together.

Disclosure: The editor of this piece consulted for the Convergent Behavioral Science Initiative at the University of Virginia, which is co-directed by the author.