Chasing the academic dream: Biased beliefs and scientific labor markets

https://doi.org/10.1016/j.jebo.2022.07.021Get rights and content
Under a Creative Commons license
open access

Abstract

We investigate whether biased beliefs play a role in the persistent demand for postdoctoral training in science. We elicit the beliefs and career preferences of doctoral students at 56 U.S. chemistry departments through a survey combined with a field experiment, in which we randomize the provision of information to a subset of respondents on historical academic placements by department. We first show that respondents have excessively optimistic beliefs about their own and their peers’ chances of obtaining a tenure track faculty position. Respondents who received the historical placement information treatment updated their beliefs about their own likelihood of obtaining a faculty position in a follow-up survey one year later, particularly those who had the most biased initial beliefs. However, we do not find an effect on the likelihood of doing a post-doc at four years post-intervention.

Keywords

Information
Biased beliefs
Career preferences
Science
Higher education

JEL classification codes

I23
D80
D84
J24

Cited by (0)

We thank Sandro Ambuehl, Michal Bauer, Albert Bravo Biosca, Donna Ginther, Stepan Jurajda, Nicola Lacetera, Mario Macis, Nikolas Mittag, Henry Sauermann, Paula Stephan, Hubert Wu, Basit Zafar and conference participants at the NBER Summer Institute, CERGE-EI, Max Planck Institute, the IGL Winter Research meeting, the Swiss Economists Abroad Association, CUNY Queens, and the University of Connecticut for helpful comments. We appreciate research assistance from Samantha Blaney. Gaulé and Vuletić Čugalj acknowledge financial support from the Czech Science Foundation (GACR grant no 16-05082S). Ganguli acknowledges support from the W.E. Upjohn Institute for Employment Research. This study has been approved by the University of Massachusetts Amherst IRB and was pre-registered in the AEA RCT Registry (AEARCTR-0003212). This paper uses confidential survey data collected by the authors, which they are unable to publicly share. The authors will provide an anonymized dataset and code to assist interested researchers with replication of the tables and figures in the paper.