Time-To-Degree Data

From: Peter Calingaert
Date: Thu, 4 May 1995
Subject: Time-to-degree data

[Editor's note: Dr. Calingaert was the Director of Graduate Studies at the time this was written, just before his retirement from the department in 1995.]

On 24 April 1995 I heard Professor Thomas Espenshade of the Office of Population Research at Princeton present findings on their study of Ph.D. students' time-to-degree (TTD). They have data on "roughly 35,000 graduate students who entered graduate school between 1962 and 1986 at eleven of the top Ph.D.-granting institutions in the U.S." (including UNC-CH). Students were in six fields: economics, English, history, mathematics, physics, and political science.

Variables considered were: university (11), field (6), cohort* year (4 periods), gender, US residence (temporary or permanent), cohort size (4), cohort's percent of temporary residents (4 groups), verbal GRE, "math" GRE (which I interpret to mean the old Quantitative), whether student in top half of those offered admission, and primary support source (fellowship, teaching, research, own funds).

(The cohort is the group of students entering the doctoral program in one department in one year.)

They measured TTD from enrollment until award of the degree (which occurs in some universities only once per year). They "[made] allowances for master's work that may have contributed" to the Ph.D. Mean TTD across the entire sample was 7.0 years. Times were less for economics, math, and physics than for the other fields. Times were less for men (6.7) than for women (8.0), and less for temporary residents (6.1) than for permanent residents (7.2). Conjectures are that field choice and, for temporary residents, visa pressures perhaps affected these TTDs. Analysis shows that half the students did (or would) finish the Ph.D. eventually.

Here's how different variables were associated with short TTD.

• High extent: field, gender, residence, cohort size (small is beautiful), percent of temporary residents (low is best), "math" GRE, whether in top half, primary support source (research is best, own funds is worst)
• Moderate extent: university (I'll withhold names), cohort year (the longer ago the better)
• Small extent: verbal GRE (this surprised the researchers)

Speaking facetiously, it appears that we should grant admission priority to applicants in the top half of those to whom we offer admission, and to nonresident males while keeping a small percentage of nonresident students. We should enroll a small cohort and support all doctoral students with research assistantships. It would also help if we were Princeton or Harvard.

Speaking seriously, analysis is fraught with difficulties, but I'll be glad to show you the many pages of "hazard analysis" tables and figures that were presented. One finding of particular interest to me was that TTD is helped much more by dissertation fellowships than by first- and second-year fellowships. Unfortunately, the latter are often needed to attract the applicants in the first place.