registered plan (Pugatch and Schroeder 2019), with analysis not specified in this plan
explicitly labelled as exploratory.
We find that email messages increased the probability that a student went on to major
in Economics. Although effect sizes were less than 2 percentage points, the magnitudes
rival the control group proportion majoring in Economics. Moreover, because the outcome
is measured by administrative data collected in the academic year following treatment, the
effects represent a durable change in revealed preference.
All effects were driven by male students, however. We find no effects among female
students, even when considering a continuous measure of interest in the major. Null effects
for female students persist across several subsamples and specifications. Among the full
sample of students, we find no effects of any message on intentions to minor in Economics.
We also find no effects of the Phase Two intervention.
Using our results on the differential effects of the intervention among male and female
students, we conduct a thought experiment. How would the male/female ratio of Economics
majors change if sending the most effective message to all Principles students became
departmental policy? In this scenario, the male/female ratio would rise from 1.4 to 2.7, an
increase of 96%. For students earning a B- or better, the implied increase is 166%. Even
under a more conservative scenario which averages over the effects of all message types,
the male/female ratio would rise by 54%. Our results should therefore sound a note of
caution about the potential for simple nudges to exacerbate inequalities within Economics.
We find several surprising results when comparing effects of different messages. In the
full sample, the varying treatment messages were no more effective at increasing Economics
majors than the placebo email. This result may be explained by low engagement with the
experimental email messages, as most students opened the emails but did not click on the
associated links. When the outcome is self-reported likelihood of majoring in Economics on
a 0-100 scale, the earnings information and AEA video messages
reduced
likelihood to
major by 2.6 and 2.2 percentage points, respectively. These negative effects were driven by
male students on the intensive margin, however, suggesting that resistance to our marketing
messages was concentrated among students unlikely to become Economics majors in any
case.
We contribute to the burgeoning literature on promoting interest in undergraduate
Economics, particularly among women and other groups underrepresented in the field. The
large scale and negligible marginal cost of our experiment—the interventions consisted of
a single email in each phase—help to understand the frontier of informational nudges to
promote undergraduate Economics. We complement Bayer, Bhanot, and Lozano (2019),
who test similar messages among incoming students at liberal arts colleges, by studying
introductory Economics students at a less-selective public university. Both studies find
positive effects of around 1 to 3 percentage points, with no significant effects for female
students. Unlike Bayer, Bhanot, and Lozano (2019), however, we find positive, precisely
estimated effects for male students, whereas their effects are statistically distinguishable
