On the Search for Talent in Academic Labor Markets: Easily Observable Later-Graduate Study Outcomes as Predictors of Early-Career Publishing, Placement, and Tenure
Hilmer, Michael, Hilmer, Christiana E., Economic Inquiry
Identifying academic talent is an important concern, both in terms of graduate school admission and persistence and in terms of tenure-track hiring, promotion, and tenure. Unfortunately, many important characteristics foretelling academic success are not directly observable, leaving admission and hiring committees to ask "What information credibly signals otherwise unobservable economics talent?" (Grove and Wu 2007. 506). A significant literature seeks to answer this question by identifying observable factors that are correlated with the unobservable determinants of future academic success (including a prominent session at the 2007 ASSA annual meetings). Despite the quality of these efforts, the research has generally agreed that
there is not an easily recognizable star profile or single path to success for an economics graduate student (Athey et al. 2007. 518).
The inability of prior empirical work to identify an easily recognizable star profile is unsurprising for a number of reasons. First, previous studies (Athey et al. 2007; Grove, Dutkowsky, and Grodner 2007; Grove and Wu 2007; Krueger and Wu 2000) have focused on measures of pre- and early-graduate school performance, such as undergraduate institution, GRE scores, application references, admission committee rankings, and performance in first-year core courses, all of which are likely indicators of a student's potential for career academic success. Second, previous studies have analyzed samples of applicants to and/or enrollees at a limited number of very top U.S. Ph.D.-granting programs in the United States (generally between one and five top five programs). (1) Third, previous studies have focused mainly on graduate school persistence (Athey et al. 2007; Grove, Dutkowsky, and Grodner 2007) and initial job placement (Buchmueller, Dominitz, and Hansen 1999: Krueger and Wu 2000; Stock and Alston 2000) and not on longer-term outcomes, such as career publishing success, tenure, and later-career placement.
The locus on experiences that occur so early in the graduate education process is unfortunate given that (1) students make a series of very important decisions during their later-graduate studies, (2)the outcomes of several of those decisions are easily observed, and (3) the outcomes of those decisions are likely much stronger predictors of the student's potential for career success than the undergraduate and early-graduate study signals previously examined. On the basis of these facts, Hilmer and Hilmer (2009) propose that the relative prominence of a student's dissertation advisor may provide information beyond the student's Ph.D. program as to the student's likelihood for early-career publishing success. Their empirical analysis of a large sample of Ph.D. recipients from top 30 U.S. economic programs in the early 1990s demonstrates that holding Ph.D. program tier constant, the relative prominence of a student's dissertation advisor is significantly correlated with both the number and quality of articles published in the student's early career. Accordingly, in-sample predictions based on their estimates suggest that students from lower-ranked programs working with prominent advisors can often be expected to out-publish otherwise similar students from higher ranked programs working with less prominent advisors.
Extending this logic, we use publicly available data sources to identify two additional easily observed later-graduate study signals that we believe likely to be more strongly correlated with career academic success than the relative prominence of a student's dissertation advisor: (1) whether the student had a paper published or accepted for publication while in graduate school and (2)whether the student coauthored a published paper with his or her advisor. We start by repeating Hilmer and Hilmer's analysis (with the addition of our two new proposed signals) for a significantly larger sample of students receiving their Ph. …