Thread by @jnaecker

Yesterday I had the pleasure of seeing @Richwpatt present at the @RadySchool econ/strategy seminar.

Rich presented the incredibly clever paper "Timing is Everything: Evidence from College Major Decisions" with Nolan Pope and Aaron Feudo.

drive.google.com/file/d/1JeTxwy…

1/10

They examine major choice among cadets at West Point, where:

(1) students must declare a major in the first semester of their sophomore year

(2) the freshman and sophomore curriculum is completely fixed, and the semester that some course are taken is assigned at random

2/10

Thus we know that the timing of taking those courses is not endogenous.

The headline result: being in a course in a given field when declaring one's major doubles likelihood of majoring in that field.

3/10

This "first-semester bias" is persistent: half of the effect still present at graduation, despite the fact that students can costlessly switch majors in the next semester by filling out a form.

The bias is also present across all student demographics.

4/10

My favorite part of the paper is the careful consideration that @Richwpatt and colleagues give to the possible mechanisms.

5/10

It could be that students are rationally responding to new information they gather through experiencing the course.

But the authors find the first-semester bias even for students who have bad experience in that semester (eg bad grades or a bad instructor).

6/10

It might also be that students are being exposed to the field for the first time, and they didn't know that it existed before the semester. But the bias is present even for students who expressed interest in the major before matriculation, or took AP classes in that area.

7/10

The authors' preferred explanation consists of two biases working together:

(1) Availability bias drives the initial choice: student conflate recent experience with the topic with enjoying it

(2) Status quo bias explains why students don’t change major in later semesters

8/10

The implications of this behavioral pattern are huge: Not getting into a class one semester can vastly decrease the probability of majoring in that field, and lead to massive differences in lifetime earnings.

9/10

For another perspective, see with recent writeup in the WaPo (which also discusses a similar paper by @Richwpatt and others on attribution bias in major choice):

washingtonpost.com/us-policy/2019…

And here is Rich's website: richardwpatterson.com

10/10