Research

Some strong opinions, weakly held

Behavior change interventions can transform businesses, governments, and everyday lives. But reality often falls short of academic theories.

The problem? We don't know what we know. We don't know what generalizes, or even agree on what generalization really means.

Better cumulation requires better coordination. Let's invest in more open, more integrative methods, together!

Do nudges really work?

During my years as a behavioral science consultant, I noticed growing skepticism from my clients and colleagues. Ted Talks, pop science books, and even academic abstracts promised powerful results from small, unobtrusive changes to marketing campaigns and employee policies. Yet reality fell short. Effects were often small and inconclusive, while the investments required to design and deploy even a quasi-experiment to detect them were considerable. Reports of replication challenges in academia only deepened our doubts about the field.

I left industry for academia to tackle this skepticism head-on.

My research seeks to answer two questions:

1. Which behavior change interventions work, for whom, and under which conditions? 

2. How can we improve our research methods to better answer these kinds of questions—questions of reproducibility, reliability, and generalizability?

I believe the insights produced by psychologists and behavioral economists contain considerable potential value for managers, marketers, and policymakers. My goal is to  help evolve how we produce, organize, evaluate, and document those insights—in order to calibrate expectations and build a more cumulative science of behavior change.

Put simply, we should know what we know.

My work combines theories from behavioral economics and judgment and decision-making with methods from meta-science and computational social science—an interdisciplinary combination that I believe is crucial to make progress on the above two research questions. Randomized controlled experiments lie at the heart of this approach. I run them in the lab or—more often—with field partners; I systematically label, combine, and analyze them; and I forecast their collective results through prediction surveys and machine learning models. 

Across my research, I am committed to open science practices—such as pre-registering studies, posting data, code, and materials—and I am particularly excited about building tools that enable academic and industry researchers to build on prior work.  Further, I believe strongly in building field partnerships beyond the United States, and I am grateful to have a growing network of partners in Latin America.