Donald Green on Field Experiments, Voting, and Making an Intellectual Contribution

Donald Green’s career, by his own telling, did not follow a straight line. It emerged through what he calls a series of “bumbling steps”: an undergraduate fascination with political philosophy that gave way to empirical research, collaborations that opened new questions, and an eventual realization—years into his career—that randomized field experimentation would become his intellectual home. Looking back, what appears coherent on paper was anything but in real time.
In this conversation, Green reflects on how academic paths actually form: through chance decisions, intellectual curiosity, and the influence of collaborators with sharply different perspectives. He explains what graduate school really demands beyond credentialing, why writing—not teaching—is the core labor of academia, and how randomized experiments transformed political science by testing assumptions that observational data had long treated as fact. Along the way, he discusses moments when his own research overturned his expectations, from voter turnout and education to the limits of persuasion.
The discussion also turns to early-career habits that matter more than raw intelligence—deep reading, intellectual breadth, and a willingness to have one’s ideas challenged—and to the risks of pursuing a collaborative, curiosity-driven research agenda without a fixed plan. Green’s advice is simple but demanding: surround yourself with people who argue forcefully, love the work enough to endure rejection, and remember that careers often make sense only in retrospect.
​
—Donald P. Green is the J.W. Burgess Professor of Political Science at Columbia University and a pioneer of randomized field experimentation in political science, with research spanning voter mobilization, persuasion, discrimination, and experimental methodology.

“Bumbling steps,” collaboration, and the real through line
Ben Wolf: Thank you so much for joining me. To begin, if you had to describe your career path as a single problem you’ve been trying to solve, what is it—and what drew you to that problem?

Donald Green: I don’t think there really is a single substantive question. My career path is more like a series of bumbling steps with no particular direction that eventually reached a moment of epiphany—but it wasn’t part of an ex ante plan.
I went to Berkeley for graduate school for no really sensible reason—they didn’t have a foreign language requirement, and my brother was already going to Berkeley. I thought I was going to study political philosophy, and I did study political philosophy, but I ended up making an abrupt turn toward empirical work.
Now I look back and say: I never actually wrote anything in political philosophy. I only wrote—in the guise of being a running dog of empiricism. There have been lots of twists and turns. One constant is that I’ve really enjoyed the social aspects of my profession. It’s fun to learn from other people and collaborate.
The most exciting moments in my career have been learning lessons from scholars with very different backgrounds. To the extent I’d give advice: surround yourself with interesting people who can argue forcefully for a new point of view. Even if you don’t embrace it, you’ll come to grips with it in a way that enriches your intellectual experience.

Graduate school: worth it, and what you’re really signing up for
BW: You mentioned you went to Berkeley for what felt like fairly arbitrary reasons. How do you value graduate school today? It’s incredibly expensive, and it’s time out of the workforce. If someone wants a path similar to yours, what trade-offs should they be considering?

DG: The fact that I went to Berkeley was arbitrary, but the idea that I would go to graduate school hit me as an undergraduate. I was taking canonical pre-law classes and not enjoying the law part very much. But when I interned with the House Energy and Commerce Committee in 1981, I discovered legislative politics—and that was an eye-opener.
The political director said, “Okay, Green, you can do scut work and answer constituency mail, or you can shut up and follow me around all summer and be a fly on the wall.” Of course I chose the latter. When I came back, I realized: this is what I want to do. I love the idea of exploration.
From that point on, I watched professors differently: how do they do what they do? How do they talk?
I went to graduate school straight out of undergrad. I didn’t know anything about anything. I was 21 when I applied—the youngest in my class. Many people I was in graduate school with had worked, done research, or had real jobs outside academia.
In some sense, the fact that I never left academia gives me an odd perspective. But I learned a lot from them. When I studied for tests, I thought the objective was to get an A. It became apparent that was a juvenile undergraduate viewpoint. The real objective is to make an intellectual contribution to a field. When you set your sights on that, you put a different kind of effort into your work—you broaden yourself and deepen yourself. That was important.
When I talk to my own students, I ask them to think hard about whether they really love doing this. It’s one thing to want to do it—you need to love it to get over the parts that are really hard, if not painful.
It’s no fun to have your work rejected and criticized. It happened to me today, and I thought: even after forty-plus years, it still stings when your work is roundly criticized.
And the thing we’re actually doing in academia—though it looks like we’re instructors—is writing. Writing is incredibly difficult to do well. So two things: do you love it, and are you a good writer? Do you love writing? Because it’s hard work.

What experiments can reveal that observation can’t
BW: You helped normalize randomized field experiments in politics. What do experiments reveal that observational data can’t?

DG: The range of things that can be studied experimentally is narrower than what can be studied observationally—partly for practical reasons, partly for ethical reasons.
But if we use observational research designs to study cause and effect, there will always be a residuum of uncertainty. There might be unobserved variables that confound the apparent causal relationship between an intervention and an outcome.
For that reason, political science—and other fields—shifted dramatically toward experiments, or designs that resemble experiments. It’s very hard to build theory on a foundation of “facts” that may not be facts.
Looking back on decades of field experimentation, things that were taken as facts didn’t stand up to scrutiny when subjected to experiments. That applies to mobilization, persuasion, and other policy principles involving costs, frictions, and resources.
It’s not that the theories were stupid. They were smart theories. The question is: did they point people in a productive direction? Often the answer was: not really.
One thing we learned is that theorizing appropriately is much more difficult than it looks—especially given how context-dependent a lot of what we study is.

Early-career habits: what matters beyond raw talent
BW: You mentioned earlier the requirements for entering academia: loving the work, loving writing. If a young scholar has those, what habits matter early on beyond raw intelligence?

DG: Read deeply enough to understand the intellectual pedigree of your field, especially the area you’re studying.
When I think back on my very best students, one thing that distinguished them is they could have a conversation not only with their peers but with people across multiple academic generations. Their reading was wide enough to give them deep perspective.
You could see it in their writing. In the introductions to their papers, they could summon ideas that would be unknown or foreign to many counterparts. So: being a deep and perceptive reader can separate two otherwise equally intelligent people.

A paper that changed his mind
BW: What’s a paper you wrote that changed your thinking—where you went in with an assumption that didn’t survive the research?

DG: It’s a little embarrassing to talk about your own work as changing your mind, but one thing that’s happened is: I’ve gone into an experiment thinking it would come out one way, and it came out the opposite.
A good example is a paper with Rachel Milstein Sondheimer on the effects of education on voter turnout.
There’s a massive cross-sectional correlation in every observational study in the U.S. between educational attainment and voter participation. I thought: that’s got to be spurious. There are lots of reasons to think factors other than education per se might explain the correlation.
So we looked for opportunities to study randomized experiments—or very close to randomized experiments—where there was an exogenous, in some cases truly random intervention that raised educational attainment in the treatment group versus the control group. The question: when they became adults, did the treatment group vote at higher rates?
I thought: no way. But actually, three for three—all showed a turnout effect, which I did not expect. That’s a good example where intuitions go one way, but the facts go another, and it changes your mind.

Career paths that only cohere in retrospect
BW: A lot of people’s careers look linear on paper, but they’ll admit they were lost in real time. When you were trying to figure out your path, what did you look for to stay on track? Did you know, or did it only make sense in retrospect?

DG: In my case, I did not have a clear intellectual agenda at first. One manifestation is that I worked on all sorts of projects on unrelated topics—which I wouldn’t recommend to people on the hunt for tenure. It’s risky. It worked for me, but I wouldn’t generalize from it.
I arrived at Yale in 1989 without a clear set of things I would study—maybe campaigns and elections, maybe public opinion, maybe methods. I worked with a political theorist on a book about rational choice theory, which sent me down a different direction. I wrote a dissertation on self-interest and political and economic behavior.
Then I became increasingly interested in discrimination, prejudice, and hate crime—another direction.
My problem is that I can get interested in almost anything, and I enjoy the social aspects of collaborative work. If the right people come along and invite me into a collaboration, I’ll go in that direction. That’s a risky strategy.
But years later—before I got to New York—I started working on randomized field experimentation. That was in 1998. And once I did that, I realized: that is my calling. That’s what I want to do for the duration of my career. But it wasn’t according to a plan.

What’s interesting now: persuasion and durability
BW: What questions interest you right now?

DG: I’m increasingly interested in whether interventions designed to be persuasive actually work—and when they do, whether they endure.
Beyond that: to what extent is it possible to change people’s minds not only about a specific proposition, but a broader suite of opinions? Is it possible to have a transformative persuasive intervention, as opposed to an aerosol-like effect that wears off quickly?

BW: And what have you found so far?

DG: I’ve done a lot of experiments—in the West and in the Global South. I’ve seen many instances where dramatization—narrative dramas—can change views in a persistent way. But those effects are fairly limited to the specific things modeled in the dramas.
So one question is: is this because my experiments have had insufficient dosage? If I studied how people absorb messages over a longer period of time, with more and more episodes, would I find stronger effects?
And another question: to what extent are pedagogical shows influential in ways that go beyond what we ordinarily appreciate—because our studies aren’t capturing everything people take in? Or is it the case that they’re entertaining and people follow the characters, but at the end of the day they’re not transformed?

Closing book recommendation
BW: Professor Green, I’ve really appreciated your time. I want to close the way I typically close Pathway Blog interviews: for a student interested in following a path similar to yours, what book would you recommend, and why?

DG: That’s a good question. There are so many great books, and it’s hard to pick one. Having just taught a great books course, the problem with reading one is you want to read them in conversation with one another.
In some ways, you want to pick one that will get you fired up—with objections—so you have to reflect on why you disagree. For example, if you want to get riled up, you could read Achen and Bartels’ Democracy for Realists. Not because you’ll be nodding along the entire time, but because you’ll have to reflect on why, and to what extent, you disagree.