When critics of graduate education talk about what's wrong with it, their poster child is someone who has been working toward a Ph.D. for 8 or 10 years with no dissertation defense in sight. Experts have talked about various ways to help such graduate students. Last week, Yale University announced a plan  to re-examine all of its Ph.D. programs -- with an emphasis on what happens in the second through fourth years -- based on the belief that what happens in those years may well determine whether a student can wrap up a doctorate in six or seven years.
"The people who are getting into trouble finishing, it's not because of what they are writing in year six. It's because they haven't started writing until year four and a half," said Jon Butler, dean of Yale's Graduate School of Arts and Sciences. By focusing on years two to four, he said, Yale hopes to be sure that departmental policies on coursework, qualifying exams, picking a dissertation topic, and mentoring all serve to make sure that by the end of that period, someone is launched toward the dissertation.
The move by Yale -- which has many top Ph.D. programs -- comes at a time of increased discussion about how to speed up the completion of Ph.D.'s,  and experts said that the approach the university is taking could be influential. "This is hugely significant," said Debra W. Stewart, president of the Council of Graduate Schools. "The Yale Graduate School is putting this in just the right context," she said, noting the emphasis in the effort both on time-to-completion and on recognizing the differences among disciplines.
Butler said that it was essential that the reviews take place department by department (Yale has more than 70 that award Ph.D.'s). "The programs are very idiosyncratic," he said, even within the humanities or social sciences or natural sciences. And at the graduate level, "students live out their lives" in the department, not the university as a whole, he said.
Each department has been asked to conduct its review this fall.
While departments may vary widely, Butler said that his goal is to have Ph.D. students feel some certainty about the length of time a doctorate will take, much the way Yale undergraduates can feel with some certainty that they can finish in four years. Currently, Yale does better than most institutions, with humanities doctorates (which in national comparisons  tend to take longer than other disciplines) finishing in just under seven years on average, but Butler said he would like to see that figure closer to six.
Butler said that the idea of having each department conduct its own review was based on the success of a recent analysis by Yale's political science department, which following its review announced a series of changes in its graduate program.
Stephen Skowronek, a professor who just completed a term as graduate director in the department, said that several changes focused on the student-adviser relationship should reduce time-to-completion rates. For example, he said that the department will now require -- on a set schedule -- a series of meetings between students and advisers in their third year. That's the year when course work is wrapping up, teaching assistant duties may be on the increase, and there "isn't a lot of structure," resulting in some students getting off track, Skowronek said.
The department is also now going to work harder to make sure good adviser-advisee relationships are established in the first two years. Skowronek said many graduate students tend to hold back on establishing a close relationship with an adviser until the students have a dissertation concept worked out. This can mean that early on, students "don't have someone helping them plan an intellectual agenda," he said.
While all students have advisers in theory, the department is now going to communicate to students that it is more important for them to have a good advising relationship right away than to worry about whether that will be the person to lead a dissertation committee. "We're telling people that we won't be offended if they change later," Skowronek said.
Beyond advising, the political scientists agreed on other changes in Ph.D. education as well. From now on, a professor will be appointed -- much as graduate directors are currently appointed -- to focus on preparing students for the job market. "We want to integrate students into professional life sooner," he said.
The department also reworked the way the way students have to demonstrate competence in the various subfield of political science (Yale currently counts eight). Instead of having to pass comprehensive exams in three fields, students will now be able to demonstrate competence through coursework in two fields while taking comps in two others.
Different disciplines will handle such questions in different ways, Skowronek said, but there is a common issue as well. "This is about how to balance specialization and broad expertise in graduate education," he said.
Stewart, of the Council of Graduate Schools, said she saw the Yale effort as emblematic of the work many graduate institutions (including Yale) are conducting through the council's Ph.D. Completion Project.  She said that with national debates about accountability in higher education capturing headlines, she was struck by the way a decentralized approach was working, as graduate schools were setting targets for review but leaving much of the decision making in departmental hands.
"These efforts are going to work because what we are seeing is highly congruent with our own cultures," she said.