A Q&A with Professor Stefanie Fischer
By Eduardo Zambrano | MS Quantitative Economics Program
This fall, I sat down with Professor Stefanie Fischer, who obtained a Ph.D. in Economics from UCSB in 2015 and has been an Assistant Professor of Economics at Cal Poly since. She teaches Introductory Economics, Labor Economics, and Econometrics. As we spoke, we covered her research, which studies peer dynamics in economically motivated scenarios. We also discussed civic engagement, data analysis, and how empirical research with a focus on inequality can have a real-world impact. Read on for more.
EZ: I see that your paper, “Coordination and Contagion: Individual Connections and Peer Mechanisms in a Randomized Field Experiment,” was just accepted at the Journal of Public Economics. Congratulations! Can you tell us what it’s about?
SF: Thank you! Yes, so to provide some context for the paper: for quite some time, economists and other social scientists have been interested in understanding how one’s peers influence economically meaningful outcomes. This paper makes a technical contribution to the peer effects literature which, to the best of our knowledge, had not been previously identified.
What we do is estimate spillovers (peer effects) between a treated individual and another treated individual, where the previous literature has only been able to estimate spillovers that arise between treated and control groups. Our approach allows us to isolate each spillover separately.
There are many settings in which there may be no spillovers present between treated and control individuals, but where there might be between the treated.
For example, safety net programs (i.e., SNAP or Head Start) are often concentrated among lower income individuals. And if lower income households are mostly interacting with other low-income households who are also recipients of such benefits, we may miss important spillovers if we don’t consider those that arise between the recipients of the benefits (the treated).
So in this paper we run a field experiment in a large university dorm where we financially incentivize a random subset of individuals to go to the university gym. In the first stage, we elicit the friend network so we know how every person is connected and to what degree. In the second stage, we randomly assign the treatment to 60 percent of the subjects. The treated subjects were told that if they visited the gym eight or more times in the 30-day period they would earn $60.
The randomization of the treatment serves two purposes: First, it induces random variation in exposure to the treatment, identifying the direct effects—that is, the effect that operates through the private incentive. Secondly, it induces random variation in exposure to treated and untreated peers, which allows us to estimate the spillover effects. Crucially, we can estimate the effect of having a treated best friend on control and treated subjects.
We find that the financial incentive, not including any spillovers, increased gym visits by about four, if you’re treated with a control best friend. We find however, that if you’re treated with a treated best friend, you visit the gym one additional time. Another way to say that is that the spillover between treated and treated is 25 percent of the direct effect of the treatment. Interestingly, we detect no spillovers between treated best friends and control subjects. It is worth noting that the observed peer effect, a spillover from treated subjects to treated subjects, would not have been identified by a standard approach, where only spillovers from treated to control are identified.
So you might wonder what is the mechanism underlying this spillover? We consider three common mechanisms put forth in the peer effects literature: The first is coordination. The second is imitation/conformity. And the third is information.
We find strong evidence in favor of coordination, where coordination may include a commitment mechanism or complementarities. Specifically, we find that subjects with treated best friends make, on average, one more simultaneous visit with their best friend than their counterpart. This suggests that the entire spillover is operating through this coordination channel.
EZ: So what’s next?
SF: While this paper is finally complete, we also have a follow-on project. We implement a similar field experiment as the gym experiment, but we incentivize students to register to vote, and then to vote. Our goal with this project is to understand the role of peer effects in civic engagement. One thing that is neat is that we have merged the experiment data with the universe of California administrative voting records. We’re still working on analyzing the results of the experiment. Stay tuned for a draft soon!
EZ: On your website you state that your research employs quasi-experimental techniques and field experiments to address policy-relevant empirical questions with a focus on inequality. How do you see the empirical tools you use in your research being of value to economists working in different capacities in companies, the government, and NGOs?
SF: The tools I use in my research are widely applicable. I use various research designs, whichever one is best for the question I am trying to answer, with the general goal of nailing down a causal relationship. I think a well-executed causal analysis is valuable in lots of contexts.
EZ: At Cal Poly you teach Econometrics and Labor Economics. In what ways is teaching these courses useful for the kind of research that you do, and in what ways the research that you do is useful for teaching these courses?
SF: Teaching Graduate Labor Economics has been useful for my own research. It helps me stay current. I often assign new papers or working papers that I want to read, which are also relevant to the course material. Sometimes through this I get new ideas for projects. I also gain a better grasp of the literature and methods that are on the cutting edge of my field (Labor Economics).
Teaching introductory undergrad econometrics is fun. I love when students have that moment in class when they realize what a useful tool applied econometrics can be!
EZ: Anything else you would like to tell us?
SF: Thanks for your interest in my work!