Prisoners of the Ivory Tower

    “How much longer do you have?” It’s the question every graduate student dreads. Whether it’s parents, friends, or new students, nobody seems to realize how much that question hurts, or, for that matter, how hard it is for graduate students to pin down the answer. Real statistics are not well-advertised, so most students don’t know the average degree time, or even the percent of students who actually graduate. Besides this, the distribution seems so broad, it is hard to know if they will follow in the footsteps of the legendary tenth year, or replicate the feat of the whiz that finished in three-and-a-half years and already has a job.

    Although many people have the vague idea that a PhD is a roughly five-year endeavor, the actual time required to complete a doctoral degree is almost always longer. In 2003, the median number of years between enrollment in graduate school and completion of the doctoral degree (elapsed time to degree, or ETD) in all fields in the US was 7.5 years.1 The national median in the physical sciences (including mathematics and chemistry) was 6.8 years; in engineering it was 6.9 years, and in the life sciences it was 7.0 years. The typical time to degree for the humanities and social sciences is longer, and nationwide, the length of all graduate degrees increased steadily during the late twentieth century. The median registered time to degree (RTD or the time a student is registered in a graduate program) in the United States for students completing their degrees in 1971 was 5.7 years, but by 1991 it had climbed to 7.0 years.2, 3

    At Berkeley, the Graduate Division keeps track of each student in a database known as the “Monster” file, which goes back to the 1960s. Berkeley’s average RTDs are shorter than the national averages, but still nowhere near five years. In 2004, the average RTD in life sciences was 6.1 ± 2.3 years; in physical sciences it was 5.7 ± 2.0 years; and in engineering it was 5.8 ± 1.8 years. Mary Ann Mason, the Dean of Berkeley’s Graduate Division, expresses the concern many feel about how long it takes to get a degree, noting, “It’s expensive in time and money for the student and for the University if the time line is too long.”

    Of course, the story these statistics tell isn’t as clear as it appears at first glance. It is difficult to incorporate data about slow finishers and students who drop out. Many databases only track students who complete their degrees, making it difficult to measure attrition from doctoral programs, which is itself a severe and complicated problem. Also, the length of graduate programs is often reported as the time between the awarding of the bachelor’s degree and the awarding of the doctoral degree (total time to degree, TTD)—roughly three years greater than RTD. Even after accounting for these factors, the statistics can be interpreted in different ways. The 1988 Survey of Earned Doctorates conducted by the National Opinion Research Center (NORC) reported that TTD had increased by approximately 30% over the preceding twenty years. But a subsequent study found that the decreasing number of PhDs awarded over this time created a statistical bias and possibly more public concern than the issue may have warranted. Calculating time to degree based on entering, rather than graduating, class showed a more accurate—and still significant—increase of 10% over the same twenty year period.3

    Happily for most, the trend has begun to level off, particularly for students in the sciences and engineering. According to the 2003 national survey,1 the median TTD was 10.1 years, or three months shorter than in 1998.

    How long should it take?

    So how long should a PhD take? The Association of American Universities was formed in 1900, and by 1916 was already questioning the ideal duration for post-graduate study, proposing three years as an appropriate length. This number had not changed much by 1964, when the Association of Graduate Schools and the Council of Graduate Schools issued a joint statement that a PhD program should take three, or at most four, years.3 This sounds great to today’s graduate students, who have had to come to terms with much longer degree times. This author’s informal survey of current Berkeley students in one graduate program showed that most think five years is the ideal RTD and six years is too much. Mason posed five years as the ideal time to degree in the sciences and engineering, and six years in the humanities and social sciences.

    Suzanne Lee and Bosun Min, graduate students in MCB, are treated like local wildlife.

    Suzanne Lee and Bosun Min, graduate students in MCB, are treated like local wildlife.

    Many graduate students have heard that doctoral programs in other countries take years less than those in the United States. The US system was originally modeled after Germany’s, albeit with a more structured program and curriculum. The German degree is designed to take around three years, though time to degree has increased there, too.4 In England, the Doctor of Philosophy degree (DPhil) requires a minimum of three years of research. According to Oxford University’s web site, DPhil students must complete a thesis based on research, “of a kind which might reasonably be expected of a diligent and competent student after three or at most four years of full-time study.” This sounds attractive compared to the US system, but it is difficult to directly compare the systems since undergraduate education is so different; for example, German and British undergraduates focus on a specialty immediately.

    But according to Dan Fletcher, assistant professor of bioengineering, who received a DPhil from Oxford as well as a PhD from Stanford University, there can be “more efficiency and less frustration in British versus US PhD programs.” He notes that the Oxford program was more structured, with a progress assessment after the first year involving both an oral exam and a plan for finishing thesis research in the remaining two years. In the US system, he says, it is possible to work on your degree until “every last ounce of interest in pursuing this field has been squeezed out of you.”

    Winners and Losers

    If everybody thinks shorter is better, why does time to degree keep increasing? One reason is obvious: the longer you’re there, the more you learn, the more research you do, and the more productively you do it. Thus, adding a year on the end of a graduate program can significantly increase students’ output in terms of research and publications, benefiting both them and their advisors. Of course, students also have reasons for wanting to get out quickly, but surprisingly, it is departments and institutions that may have the biggest interest in getting students out the door. Berkeley’s job is to produce graduates, particularly those who go on to notable careers. The faster they can do it, the more they can produce. Mason explains that time to degree is “often less important for the graduate student than it is for the institution…there’s a good deal of expense and effort in training graduate students and the longer they stay, the tardier the payoff is in the sense that they are moving into the academic pipeline and continuing on with their careers.”

    Without a doubt, long average TTD has broad repercussions. The recent summary of the Survey of Earned Doctorates by NORC noted its impact on “the faculties and administrations of the degree-granting institutions, as well as national public agencies and private organizations that support doctoral study.” This wasn’t news to the University of California. A feared shortage of PhDs in the 1980s led the California State Senate to commission a study of the time to degree in the UC system by Maresi Nerad and Joseph Cerny, who was then Berkeley’s Vice Provost for Research. More recently, in 2001 the UC convened a Commission on the Growth and Support of Graduate Education. Based on the anticipated undergraduate population and the needs of industry in California, the commission predicted a need for 40,000 new faculty by 2010. Consequently, they set the goal of 11,000 new graduate students in this decade. California is one of only five states in which graduate enrollment has declined — by 2% in the last decade — while nationwide enrollment has increased. Among the commission’s proposals to reach its goal was one to “expand current efforts to reduce the time graduate students take to complete their programs,” by monitoring progress closely, providing support services, and increasing student financial support.

    Cerny, a professor of chemistry and former dean of graduate studies at Berkeley for fifteen years, feels that Berkeley’s current graduate program lengths in the sciences and engineering are reasonable for a top research institution, but are a factor in what he believes is the larger problem of the increasing time to the first job for scientists and engineers. Cerny was recently part of a committee to study education in his field, nuclear science. Their findings included the following:

    “The median registered time from bachelor’s degree to a PhD in nuclear physics or nuclear chemistry has been 7.0 years over the last five reporting periods (1998–2002). Seventy percent of these PhDs then take one or more (almost mandatory) postdoctoral positions lasting an average of 3.3 years. Therefore, ten-plus years pass before the ‘typical’ nuclear science PhD has a first job. This is too long…We believe that the time to the PhD should be shortened to five and a half or six years.”5

    Part of the problem, according to Cerny, lies in inadequate career advising: “Departments ought to be doing much more career advising as to what the typical jobs actually are.” Giving career guidance is not a high priority for many busy professors, nor are they often familiar with the many career paths their students could follow.

    A 1995 report entitled “Reshaping the Graduate Education of Scientists and Engineers” noted the 30% increase in time to degree in some fields and stated: “Spending time in doctoral or postdoctoral activities might not be the most effective way to use the talents of young scientists and engineers for most employment positions. Furthermore, because of the potential financial and opportunity costs, it might discourage highly talented people from going into or staying in science and engineering.”6

    Mason also emphasizes the importance of decreasing the time to first job. She noted that, more than graduate student time to degrees, “A bigger concern in this regard are the post-doc appointments, which tend to get longer and longer all the time. In the biological sciences it’s not uncommon to have a five year postdoc before your first job…That’s just elongating the experience for everyone, which can be very demoralizing, to be honest.”

    Source: Digest of Education Statistics (2003)

    Source: Digest of Education Statistics (2003)


    Source: UC Berkeley Graduate Division

    Source: UC Berkeley Graduate Division

    Their fates in whose hands?

    There are a variety of opinions on who or what controls the time to degree. Asking a small sample of current PhD students at Berkeley whether the time they would take to complete their degree was within their control, responses varied from “it depends on my advisor” to “[it’s] managed by the individual.” The encouragement of their advisors and agreement on a research plan are both contributing factors. In general, students earlier in their programs felt they had less control, while students closer to graduation mentioned the effects of unexpected problems with experiments and difficulty maintaining focus. One student said, “There [were] way too many interesting projects that I wanted to do. A good portion of the projects aren’t thesis-related, but just too good not to investigate.”

    When asked how students could influence their own completion times, Cerny advised that students “get into research as early as possible…figure out early what you want to do and start it.” He noted that in the Department of Chemistry, most students join a lab by the middle of their first semester. This has been true for the last half century, and, says Cerny, “Chemistry’s time to degree has been around 5.3 years for the past forty years.”

    Good mentoring was described by Dean Mason as one of the most significant factors in graduate students’ academic experience. “For the graduate student, having a really good mentoring experience is probably the most important thing. And good mentors want to get them out…which means they’re really concerned about the welfare of that student as well.” A 1996 study of former doctoral students supports Mason’s belief, finding “that students’ aspirations to the doctorate are most affected by the relationship they have with their individual faculty research advisor.”7 For both men and women who left Berkeley before finishing their degrees, 54% cited a “lack of guidance from my advisor” as a factor in their decision to leave.7,8 In a multi-university study, interviews with faculty confirmed the lack of guidance, though the faculty view the problem from a very different perspective: their most common advising attitude was one of an open door without making any overtures.3

    An important influence on time to degree is the unique set of mental and emotional challenges faced by graduate students. At Berkeley, 43% of the students seen at the Counseling and Psychiatric Services are graduate students, though they make up less than 30% of the student body.9 A survey by a Berkeley task force on Graduate Mental Health10 found that 45% reported emotional or stress-related problems within the past twelve months that had significantly affected their well-being or academic performance. 39% reported feeling frequently overwhelmed, while approximately 42% reported feeling exhausted “frequently” or “all of the time.”

    Though personal or individual factors may determine where each student falls within the range of times to degree, larger-scale institutional factors—outside each student’s control—can have much greater influence on average times to degree. Funding policy is one of the primary institutional means of controlling time to degree. Other things being equal, students who receive fellowships or research assistantships have higher completion rates and shorter times to degree than students receiving teaching assistantships or fee waivers or those who are totally self-supporting.10 “It’s a matter of funding structure, and if you’re structured to finish in five and half years and then you’re cut off,” says Mason, “your professor is geared to giving you the appropriate project and to working more expeditiously with you.”

    According to Mason, a good example of the influence of funding on graduation times occurred following her implementation of the Dean’s Normative Time Fellowship. This fellowship is awarded in the Social Sciences and Humanities (where times to degree are longer, and funding less available) and provides dissertation funding for students who advance to candidacy within the time limit established for their program. The program has directly decreased time to advancement to candidacy for many students because they need to make quick progress to qualify for the fellowship.

    Six institutional factors that influence time to degree were cited in the 1991 study by Nerad and Cerny: degree requirements; teaching requirements and means of evaluating progress; faculty advising and departmental guidance; student financial burden, financial support and debt accumulation; campus facilities; and the availability of professional job opportunities and placement support.11 Not surprisingly, programs that require a dissertation prospectus and an early start on dissertation research have shorter times to degree. Further, programs that evaluate students’ progress annually and provide feedback have shorter times to degree. Teaching requirements also have an effect. At Berkeley, students teaching for more than three years average one year longer to graduate than students teaching less than three years.11

    Program quality is a factor that has consistently emerged as impacting time to degree. The 1995 report, “Research Programs in the United States: Continuity and Change” found that in 1993, PhD recipients whose programs ranked in the top quartile, “typically completed their studies more rapidly than graduates of lower ranked programs regardless of field.”12 Another study found that, in general, better established programs of recognized quality and of smaller size had lower times to degree and higher completion rates. In an analysis of TTD in math and physics at three large (Berkeley, Chicago, Columbia) and four small (Cornell, Harvard, Princeton, and Stanford) universities, the larger programs averaged 6.1 years, while the smaller ones averaged 5.2 years (1967-1976 entering class). This study did not adjust for the fact that all the small schools cited were private institutions.3

    Knowing is half the battle

    The clarity with which the guidelines for completing a PhD program are set out also influences time to degree.3 A lack of information is not just frustrating for the students—it actually increases the time to degree. One survey found that only 45% of the students responding said they had a very clear understanding of the criteria for determining when they would be ready to graduate. Similarly, only 31% said that they clearly understood the length of time they would be students.13
    An informal survey of Berkeley graduate students in the sciences found that most expected to graduate after a total of four to six years, or in other words, approximately the same amount of time they thought was ideal. Based on Berkeley’s statistics, a lot of them will take longer than that to graduate.

    A little digging through departmental web sites at Berkeley usually yields one or two sentences about the duration of the program, often citing something like, “around five years.” Many programs list the “Normative Time” for their program, defined in Berkeley’s Handbook of Graduate Studies as the elapsed time students would need to complete all requirements for the doctorate, assuming that they are making adequate progress toward their degrees. But this is not necessarily the actual time required to finish doctoral studies, something only one program at Berkeley (Microbiology) states clearly on their web page for current and prospective students.

    Molecular and Cell Biology is one of the few programs in which the expectation that students will complete their degrees in five years is clearly laid out in a schedule of thesis committee meetings and progress reports, but in other programs, expectations are communicated indirectly or more ambiguously. For example, the Biophysics Graduate Group implies the length of its program by stating that it guarantees only five and a half years of funding.

    Where we go from here

    Though many aren’t aware of it, the efforts of many institutions and national bodies are focused on improving graduate education, which includes decreasing the time to degree. But Mason points out that making these improvements is like shooting at a moving target: “Things do change so these questions can’t just be resolved once. Over the last ten years we have improved a little bit. It’s always a new set of faculty and a new set of students, so it’s a never- ending [situation].” For example, at this year’s first meeting, on January 19th, the UC Regents were presented with a detailed report on the importance of graduate education to California and the UC system—an effort to draw attention back to important issues that had been neglected since the work of the 2001 Commission.
    For the individual graduate student, whether taking courses, preparing for a qualifying exam, or planning the next experiment, it can be hard to see beyond the daily grind, but taking a step back and looking at the big picture can pay off. Studies show that awareness of graduation requirements, actively seeking mentoring resources, writing a dissertation prospectus, and meeting with a dissertation committee can all help get the student out of the lab, into a cap and gown, and then onto the streets of the real world.


    1. Hoffer, S., Welch, Williams, Hess, Friedman, Reyes, Webber, Guzman-Barron, Doctorate Recipients from United States Universities: Summary Report 2003. NORC at the University of Chicago, 2004.
    2. Summary Report 1996 Doctorate Recipients from United States Universities. National Academy of Sciences, 1998.
    3. Bowen, W.G. and N.L. Rudenstine, In Pursuit of the PhD. Princeton, NJ: Princeton University Press, 1992.
    4. Nerad, M., Preparing for the Next Generation of Professionals and Scholars: Recent Trends in Graduate Education in Germany and Japan. UC Berkeley, 1994.
    5. Education in Nuclear Science: A Status Report and Recommendations for the Beginning of the 21st Century. DOE/NSF Nuclear Science Advisory Committee Subcommittee on Education, 2004.
    6. Reshaping the Graduate Education of Scientists and Engineers. National Academy of Sciences, National Academy of Engineers, Institute of Medicine: Washington, DC, 1995.
    7. Humphreys, S. The Role of Women Graduate Students in EECS at UCB. in Bridging the Gap. Carnegie Mellon University, 1995.
    8. Kamas, L., Paxson, P., Wang. A, and Blau, R., PhD Student Attrition in the EECS Department at the University of California at Berkeley. Women in Computer Science and Engineering (WICSE), 1996.
    9. Prioritizing Mental Health, A campus Imperative: Proposal for a standing Academic Senate Subcommittee Addressing Student Mental Health. UC Berkeley, 2003.
    10. Ehrenberg, R.G. and P.G. Mavros, Do doctoral students’ financial support patterns affect their times- to- degree and completion probabilities? National Bureau of Economics Research: Cambridge, MA, 1992.
    11. Nerad, M., Doctoral Education at the University of California and Factors Affecting Time to Degree. University of California, Office of the President: Oakland, CA, 1991.
    12. Goldberger, M.L., B.A.Maher, and P.E. Flattau, eds. Research Doctorate Programs in the United States: Continuity and Change. Committee for the Study of Research Doctorate Programs in the United States, National Research Council (US), 1995.
    13. Golde, C.M. and T.M. Dore, “At Cross Purposes: What the experiences of doctoral students reveal about doctoral education.” The Pew Charitable Trusts: Philadelphia, PA, 2001.

    Leave a Reply