Earning a Ph.D. in the humanities in the United States is demanding. It also takes years to complete—longer than in the sciences and the quantitative social sciences. For example, the median time from entering graduate school to gaining a Ph.D. for students who received their degrees in 2006 was 2.5 to three years longer in the humanities than in the life sciences, physical sciences, engineering, or economics.
In fact, about half of the students who graduate in the humanities take eight years or more to do so, and only about half of those who begin graduate school actually finish. The combination of lengthy “times to degree” and high attrition rates has burdened graduate programs and students and has plagued generations of scholars in the humanities.
Those problems led the Andrew W. Mellon Foundation to begin the Graduate Education Initiative in 1991. Over the next 10 years, the foundation spent a total of $85-million to enable 54 humanities departments in 10 major universities to improve the structure, organization, and financial support of their Ph.D. programs, while maintaining or improving the quality of the education that they offered. The foundation also supported a study of the project’s outcomes through multifaceted data collection.
In Educating Scholars: Doctoral Education in the Humanities, being published this month by Princeton University Press, we report the results of the study and chart what happened in the departments involved, as well as in a set of control departments. The study was based on annual data on the progress of all students who entered those Ph.D. programs, annual reports by professors, and the results of a retrospective survey of all students pursuing the degree in both treatment and control departments.
What were our key findings? The study revealed that times to degree and attrition rates are helped most by improved financial support for students, by clear statements of departmental expectations, and by advisers’ counseling timely completion of the dissertation. Specifically, we found that:
- Multiyear financial-aid packages, which spread at these institutions in the 1990s in the heated competition for first-rate students, appear to have reduced attrition early in students’ graduate careers. But to a large extent, that reduction has been offset by increases in the number of graduate students who drop out later. Put simply, multiyear financial-aid packages often have the effect of substituting late attrition for early attrition.
- A large share of the graduates—about 25 percent—completed their degrees after remaining for 10 years or even longer. One should not assume that students beyond their 10th year of graduate study in humanities fields will never finish their degrees.
- Data on attrition from departmental Ph.D. programs generally overstate its extent. Almost 12 percent of the students who left graduate school at one point ultimately received their doctoral degrees—either in the same field at other institutions or in other fields. An additional 18 percent earned professional degrees in fields like law and business.
- We found no single answer to the question of how long it “should” take to earn a Ph.D. in the humanities if the faculty’s main concern is students’ successful placement. Neither speedy completion (finishing in three to five years) nor unhurried completion (taking eight years or more) are as preferable as finishing within the two extremes. In the study, those who completed their Ph.D.'s in five or six years did no better in the job market than students who finished in seven years. But beyond seven years, the probability of getting a tenure-track position declined as degree times lengthened. We found no indication that protracted degrees make better scholars.
- Because publishing while in graduate school also relates to obtaining tenure-track positions, faculty members usually tend to encourage students to publish before earning degrees if they can finish within seven years. Similarly, professors concerned about their students’ job prospects have every reason to urge students who might take eight years or more to finish faster. That subtle distinction helps explain the modest effects we found the project had on time to degree. Moreover, the Mellon Foundation’s goal of reducing that time was not uniformly shared by faculty members, because some were skeptical that high-quality dissertations could be done quickly.
- Gender per se was not a predictor of success in graduate school. The completion rates of women in our study and their times to degree were not adversely affected by being married or having children at the beginning of graduate school. Indeed, on average, married and single women did as well as single men—although married men had higher completion rates and finished their degrees faster than single men.
- The notion that students are held back by teaching “too much” needs qualification. Serving as a teaching assistant for a reasonable number of terms—five or fewer—did not delay degree completion, but longer stints did. Also, the effects of holding a teaching assistantship depend not only on how long students serve, but also on when in their studies they do so. Our findings suggest that having a fellowship rather than an assistantship in the fifth and sixth years would increase the completion rate by more than 4 percent by the end of the sixth year.
- The Ph.D.'s in our study did not remain in their first jobs—within three years of earning their degrees, many moved to better positions. While unemployment continued to be low throughout the period of our study, the proportion of graduates who received tenure-track jobs upon completing their degrees fell during the 1990s and early 2000s. Still, the probability of moving from a non-tenure-track job to a tenure-track job within three years of graduation remained at about 60 percent throughout the period.
Such findings are complicated, yet for faculty members and academic administrators interested in increasing timely completion, the main lessons concern the financial support of grad students, the advising they receive, and the fact that neither very short nor very long periods working toward degrees are preferable. The study also revealed the benefits of having reliable information on students’ progress. Specific suggestions for institutions and their leaders include:
First, graduate students who are generously supported are more likely to complete their degrees in a timely fashion than those who are not. However, the form that financial aid takes, and when it is awarded, also makes a difference. For example, fellowships improve the probability of completion and reduce the time it takes to get a degree more than teaching assistantships do—although the differences in outcomes are not great. Summer support is especially advantageous. Other things being equal, it is positively associated with timely completion of the degree. Reallocating some fellowship money to summer support may reduce attrition and increase completion.
Second, multiyear financial-aid packages are now part of the landscape at the universities that we studied. But institutions should exercise caution in extending those awards and should monitor the progress of students who receive them—especially given the associations we found between having guaranteed support and late attrition.
Third, although money influences the progress that students make, major expenditures are not required. Informing students on what they must do to complete the degree and furnishing them with explicit timetables for meeting departmental requirements are not costly. Similarly, conducting periodic reviews of curricula and students’ progress will require faculty members’ time, but those activities may actually save time for professors and students in the long run. Students who reported little or no faculty attention during their first two years were more likely than others to leave later on.
Fourth, universities can significantly reduce attrition in students’ early years of graduate study, but attrition remains problematic even for students who have passed exams and are working on dissertations. Although students report that they spent more time with advisers while working on their dissertations than earlier in their studies, more hard thinking is needed to make advising more effective at later stages in the process.
Fifth, universities should be wary of trying to reduce attrition while also trying to decrease time to degree. Too often the goal of lowering the number of students who drop out is met only by reducing early attrition. That can increase the time to degree or cause later attrition that might not have occurred otherwise.
Perhaps the most heartening lesson from the Graduate Education Initiative is that raising awareness of how many grad students drop out and how long degrees actually take can stimulate thoughtful reviews of programs and sensible decisions to change them. Since 1990 new data and an abundance of studies on attrition and completion have motivated faculty members to reconsider the shape of Ph.D. programs, led students to demand information about program length and completion rates, and encouraged universities to identify ways to measure and increase program effectiveness. We hope that our study reinforces such efforts.
