Wharton's Judd Kessler and Clayton Featherstone discuss their research on social information.

Teach for America places thousands of recent college graduates in urban and rural public schools across the country each year. New Wharton research found that the inclusion of just one sentence in the job offer letter significantly increased the number of people who decided to accept. Business economics and public policy professors Clayton Featherstone and Judd Kessler joined Knowledge at Wharton to discuss how “social information” can affect the outcome of high-stakes decisions, such as choosing that first job. Their paper, “Can Social Information Affect What Job You Choose and Keep?,” which was co-authored with Harvard visiting professor Lucas C. Coffman, was published in the January issue of the American Economic Journal: Applied Economics.

An edited transcript of the conversation follows.

Knowledge at Wharton: Clayton, could you give us a brief summary of what you studied and a definition of social information?

Clayton Featherstone: Teach for America tries to place teachers in schools. If you’re admitted to the program, you get an email that says something like, “Congratulations, you’ve been admitted to Teach for America. You’ve been assigned to wherever. We hope you’ll join us.” That email is how they communicate that you should join Teach for America. The sentence we added to the email was, “Last year, 84% of people in your position chose to join Teach for America. We hope you will as well.” We found that one sentence was actually pretty powerful in inducing extra people to join. That sentence is a canonical example of social information. Basically, when I’m thinking about doing something, I might be interested in what others in my situation have chosen to do in the same way.

Knowledge at Wharton: Could you tell me some of the key takeaways of the study?

Judd Kessler: As Clayton mentioned, we found that adding this line increased the probability that people would join Teach for America. We were quite interested in that finding because social information had been used in many other contexts for very small-stakes decisions, such as deciding whether to donate $5 to support students at your university. Learning that many other students were willing to donate the $5 made others more likely to donate. Before we started our study — which is co-authored with Lucas Coffman, who’s now at Harvard — we had not seen that social information can work for a very high-stakes decision like your first job or when you switch careers and decide to become a teacher, as many in Teach for America are doing.

It’s not just that it affected a high-stakes decision, but also that it persisted. I tell you, “Look, all these other people joined Teach for America,” and now you’re more likely to join. We sort of convinced you to do it, but you weren’t really interested and now you’re going to drop out at your first opportunity — we were very worried that that might be the effect we saw. But what we saw is that not only were people more likely to join, they actually stayed at high, if not higher, rates. So, the people who were induced to join from our line in the acceptance letter were still there two years later.

“What are the very low-cost ways that you might increase some outcome that you care about, whether it’s reaching out to potential customers or potential employees?”–Judd Kessler

Featherstone: I think it’s interesting when you take a look at the study because it’s not just about a high-stakes decision, it’s also about a high-stakes decision where you have to continue to invest in the decision over time. It stands in stark contrast to other places where we’ve seen social information work, like whether you re-use your towel. You kind of choose to do it for one day, and that’s it. Whereas, you’re in Teach for America for two years.

Knowledge at Wharton: Were there other conclusions that surprised you in the study?

Featherstone: One of the things that was surprising was that we saw a bigger effect of the social information among those you might be able to identify as on the fence. For instance, instead of the entirety of people admitted to Teach for America, you look at the people who perhaps didn’t get assigned to their No. 1 region and asked them how they felt about this, and maybe they weren’t so jazzed about it. You actually saw that the social information had a larger effect on these guys.

The idea here is that these people are on the fence, so they’re 50/50 on whether they join Teach for America. These are the people for whom a small nudge, like the one that we used, might change their decision from not joining to joining.

Knowledge at Wharton: What are some of the practical implications of this? You said that Teach for America has added this sentence into its acceptance letters. But how can this be applied by other businesses or organizations, or even in other high-stakes situations?

Kessler: I think there’s a few key takeaways. The first one is, we had a measurable impact on the number of people who joined. Adding a line to an acceptance letter, particularly when it’s an email, is very cheap, if not free, to do. The fact that they now include the line in all their acceptance letters is an indication that at no cost they’re taking advantage of this social information nudge to get more people to join their program.

One potential implication for others is that other organizations might want to look for opportunities like this. What are the very low-cost ways that you might increase some outcome that you care about, whether it’s reaching out to potential customers or potential employees? These types of nudges, at a very low cost, can have big impact.

One of the things that Clayton mentioned earlier, this issue of the people who are on the fence being the ones that are affected by it — I think that’s one of the key takeaways for other organizations. With Teach for America, it’s easy to identify the people who are on the fence. They wanted to work in Philadelphia, they wanted to teach math, and they didn’t get a Philadelphia math job. They got something else. That lets Teach for America and us, as researchers partnering with them, really think about what might be important for this particular group of folks. Other organizations might want to do the same. When they’re thinking about how they might nudge people to behave, one takeaway is that they should be really thinking about who are the people on the fence who might be impacted by the intervention they’re considering.

Knowledge at Wharton: You say that knowing how many people accepted last year influenced people accepting in that following year. Was the social information influencing people because they thought the program was more competitive based on that? Did it make them think it was more prestigious? What is driving this?

Featherstone: That’s a very good question and, unfortunately, one we don’t have a great answer for. After the fact, we surveyed people who received our treated email and those who didn’t. We asked them questions like, “How good do you think Teach for America is as an organization? How much are you going to be able to help kids with Teach for America? How is this going to help you in a subsequent career?” We found there weren’t huge changes that our little information nudge made on those measurements.

That being said, it doesn’t mean that social information isn’t changing people’s beliefs about some quality of Teach for America. It just means we maybe didn’t hit that quality dimension, I guess.

Kessler: I think it’s hard to tease apart, for some of these interventions, exactly what is driving them. You raised a number of interesting hypotheses for why that might influence people’s decisions. Just because we didn’t find evidence of it doesn’t mean that they aren’t all active. Again, that’s something to think about when you’re trying to target the people that are on the fence, trying to understand their psychology and what they care about, and designing interventions that can influence them.

Knowledge at Wharton: Do you feel that there are misperceptions dispelled by this research?

Kessler: One misperception you might have is that you have to engage in lots of costly behavior to persuade people to do something. Again, the fact that this was an intervention that was essentially free and had a big impact on people’s decision to join means that there are opportunities for organizations to take advantage of similar-style nudges.

“When I’m thinking about doing something, I might be interested in what others in my situation have chosen to do in the same way.”–Clayton Featherstone

One additional thing to point out is that we had a hypothesis that this would work, but one of the important features of our study is that it was designed as an A/B test. So, we explicitly randomized which people received which of the letters, which allowed us to make very strong claims about the fact that it was this line and not something else that caused the applicants who were admitted to then say yes to the program.

I don’t know that we’re dispelling some misconception that organizations have, but if you thought that you could take a nudge out of the box and apply it to your organization without testing it, I would hope that someone would think twice and say, “Maybe we should do an A/B test first to make sure it has the effect that we think it will before running with it.”

Knowledge at Wharton: What sets this research apart from other research conducted on similar topics?

Featherstone: I think the big difference with our research is that it shows social information affects big decisions, not just whether you donate $5 or something like that. Also — and I think this is really what sets our research apart — it shows that it can affect a decision where it’s not just whether you do something today or even tomorrow, but whether you do something over the course of two years.

Knowledge at Wharton: What’s next for the research?

Kessler: We have really enjoyed working with Teach for America. They’re eager to learn from partners with academics like us, so that’s been really great. We’re working with them to try to see if there’s other ways we can help improve their matriculation and retention. One of the next things that we’re looking at is they have a transitional grants and loans program. If you get accepted to Teach for America, you’re going to be asked to move to a new city and get trained. There’s going to be a few months before your first paycheck comes as a teacher, so they provide grants and loans to their admits to help them ease that transition.

We’re helping them think about what is the best way that they could allocate those funds and investigate what the effect of an additional few hundred dollars is on somebody deciding that they can say yes to Teach for America, pick up and move, and stick with it. That’s something we’re really excited about because now we have an active relationship with Teach for America on research.