Chronicle of Higher Education
On the first day of graduate school, everyone
is still a success. All of the students
gunning for Ph.D.'s have lived an academic
life of achievement: honor roll, summa cum
laude, certificates, scholarships, and
In some humanities programs, only one of
every three entering students goes on to
earn a doctorate. No comprehensive national
statistics are available, but studies suggest
that the attrition rate for Ph.D. programs
is 40 percent to 50 percent.
That has been the way graduate school has
worked for years. It's about separating
the wheat from the chaff, some professors
will argue. Others may spout additional
clichs about cream rising and sink-or-swim
environments. The good students get through,
Nearly everyone involved, from graduate
deans to professors, acknowledges that Ph.D.
programs will never have the completion
rates of shorter, more clearly defined programs
like law and business schools. Some dropouts
are to be expected, since getting a Ph.D.
can often take six or seven years,
"If actual attrition is really around
50 percent, then this is a scandal,"
says Michael S. Teitelbaum, a program director
at the Alfred P. Sloan Foundation. "It's
a serious waste of resources and a terrible
waste of time and energy on the part of
Some researchers have tracked attrition
for years. Their studies don't suggest a
rise or a fall in the dropout rate. What
is changing is university administrators'
willingness to do something about the problem.
At the recent annual meeting of the Council
of Graduate Schools, in San Francisco, one
of the most talked-about sessions focused
Lewis Siegel, chairman of the council and
dean of the Graduate School at Duke University,
calls it "the central issue in doctoral
education in the United States today."
Debra Stewart, the council's president,
calls it a "wedge issue." Start
dealing with why people are leaving graduate
school, she says, and you'll fix a whole
bunch of problems.
The timing seems right as well. Ms. Stewart
says that the council has been overwhelmed
by requests for its new booklet examining
the research that has been done on the issue.
Just after the meeting, the council announced
that Pfizer Inc., the pharmaceutical company,
was donating $2-million for 10 pilot projects
to study and develop ways to stem attrition.
Why It Matters
Lots of people can't cut it in graduate
school, runs the common wisdom. That's the
nature of the beast. In his presentation
at the San Francisco meeting, Peter Diffley,
an associate dean of the Graduate School
at the University of Notre Dame, posed the
essential question: "Why care? Doesn't
it just cull the bottom part of class? Won't
solving the attrition problem just worsen
the placement problem?"
After studying 10 years of data in four
representative departments at Notre Dame,
Mr. Diffley found that those simple explanations
don't hold water. His research suggests
that there is little to no academic difference
between the people who complete their degrees
and those who drop out -- at least as measured
by their Graduate Record Examination scores
and undergraduate grades. So, ultimately,
the high attrition is a waste of time and
He also calculated that Notre Dame would
save $1-million a year in stipends alone
if attrition went down by 10 percent, because
programs would not over-enroll students
to compensate for attrition. "We don't
mind spending if there's a product at the
end," he says.
Graduate-school administrators also argue
that decreasing the number of doctoral dropouts
is the fastest way to graduate more American
and minority Ph.D.'s. Many of the deans
at the San Francisco meeting were worried
about what they called a shrinking "domestic
talent pool." In the past five years,
the number of Americans earning doctorates
has fallen by more than 8 percent. Meanwhile,
the number of foreign students on temporary
visas earning doctorates has risen by more
than 5 percent.
The most important reason to care about
attrition, most researchers agree, is the
effect it has on students' lives. "This
is tremendously painful," says Barbara
E. Lovitts, who left two doctoral programs
before finishing a third one, in sociology,
at the University of Maryland at College
Park in 1996.
Now a research scientist at Maryland, she
is the author of Leaving the Ivory Tower:
The Causes and Consequences of Departure
From Doctoral Study (Rowman and Littlefield,
2001). She saw several people who had not
completed their degrees cry during interviews
about their grad-school experiences and
the effect it had on their lives -- no matter
what their reasons for leaving.
"There is a tremendous opportunity
cost," Ms. Lovitts says. "These
are people who have never failed before
in their lives. They were summa cum laude,
Phi Beta Kappa. And for the first time in
their lives they've experienced failure.
It takes people a lot of years to get over
What We Do Know
Even though no comprehensive national studies
have been done on attrition from Ph.D. programs,
researchers still know a lot about the problem.
Many institution-specific studies in recent
decades bear out the same trends: Women
drop out at a higher rate than men. Minority
students leave at a higher rate than white
students do. Americans drop out more often
than international students. And students
leave humanities and social-science programs
at a higher rate than those in the sciences.
Researchers and deans concerned about attrition
say the first step in reducing it is to
gather and publish data on the issue. Admittedly,
those would be slippery statistics. How
do you decide who counts as a doctoral student?
What about students who entered Ph.D. programs
but left with master's degrees? And when
do you count them as gone? What if they
just took a leave for a year and expect
Chris M. Golde, a senior scholar at the
Carnegie Foundation for the Advancement
of Teaching who is one of the nation's foremost
researchers on graduate-school attrition,
tells departments "not to get hung
up on the technical part." Just come
up with a definition that can be easily
explained, and a way to gather the data.
Then keep putting those data in front of
people and ask them to confront the problem.
Many departments, deans agree, don't realize
the size of the problem until they see the
Prospective students often have to hunt
around to find information about completion
rates or attrition. If they knew that a
program had a low completion rate, perhaps
that would change their decision to apply.
But at some graduate schools, such information
simply isn't available.
Institutions that make it easily accessible
are in the minority. At the University of
California at San Diego, a comprehensive
table of completion statistics is included
in the graduate school's annual report,
easily found on the Web. Duke University's
Graduate School includes links to a wealth
of admissions, enrollment, and completion
data for prospective students on its Web
site. There you can learn very specific
information about individual departments
-- for instance, that the Ph.D. program
in literature bucks the national trend.
Of the 27 students who started from 1992
through 1995, 17 earned their degrees, and
only 6 have withdrawn.
The Selection Factor
While some students certainly leave Ph.D.
programs because they can't do the work,
deans say the problem is not usually students'
struggling to measure up. A larger portion
of the dropout total can be attributed to
grad schools' having made bad admissions
selections. That doesn't mean the students
aren't bright enough. Deans and researchers
talk, instead, about that hard-to-define
Even students who make it through the rigorous
selection process to win National Science
Foundation graduate-research fellowships
finish their Ph.D.'s at a rate of only about
75 percent. That's just a bit higher than
other doctoral students in the sciences.
At Duke, Mr. Siegel, the dean, has taken
to asking department chairmen what proportion
of their conversations with prospective
graduate students is "about informing
students rather than selling your program."
At the University of Illinois at Urbana-Champaign,
Richard Wheeler, dean of the Graduate College,
says avoiding bad matches is tough in a
highly competitive environment. "The
higher you crank up the screws on recruiting,
the more likely you are to bring in students
that don't fit with your program,"
Ms. Golde, of the Carnegie foundation,
suggests looking at how science departments,
which have higher completion rates than
humanities departments, generally select
"One reason the sciences have lower
attrition rates is that you are admitted
to be in the Joe Schmoe lab," she says.
You and Professor Schmoe "have spent
some time getting to know each other and
vet each other." That's quite different,
she says, from a student who plans to study
international labor economics but, after
doing years of course work, realizes that
there is no one in the department for him
to work with. "Why did you admit me?"
the student asks. "Why did you come?"
the department counters.
"It's like a bad dating situation,"
says Ms. Golde. "No one is taking responsibility
for the match. Instead everyone needs to
take responsibility for the match."
The Money Factor
You don't need a Ph.D. to figure out that
struggling to rub two dimes together for
the seven years it takes to get a doctorate
makes getting the degree harder. Money does
matter. But maybe not in predictable ways.
Holding a research assistantship improves
a student's chances of completion. Teaching
assistantships help too, although to a lesser
degree. Maryland's Ms. Lovitts, who studied
attrition at two research universities,
says money alone isn't enough. Students
on fellowships, for instance, do not complete
their degrees at a higher-than-average rate.
Assistantships really help, she says, because
they increase the likelihood that graduate
students will interact with other graduate
students and with faculty members. "You
have to come up on campus and engage in
the professional task of the department,"
she says. "You have to interact with
faculty. You get to interact with undergraduates.
You're far more likely to get a desk with
other graduate students, which puts you
in contact with the graduate-student subculture."
Ms. Golde emphasizes that this is another
way that the sciences are structured differently
from the humanities. In a science department,
students are in the lab from the start,
working next to undergraduates, researchers,
and professors. In English, on the other
hand, the first couple of years of graduate
school are taken up mostly with classes.
"It's just like being a supercharged
English undergraduate," she says. "It's
not anything like being an English professor."
About 10 years ago, Washington University
in St. Louis made a policy shift that administrators
credit with substantially raising completion
rates. The size of the graduate school was
changed to match the number of assistantships
that departments could support. That meant
a reduction in overall enrollment, but also
that every student was now assured of a
fellowship or teaching assistantship for
The move has cut attrition, says Robert
E. Thach, dean of the Graduate School of
Arts and Sciences. The average completion
rate is now 70 percent, and he hopes to
keep pushing it higher. In some disciplines
the change has been sharp. A decade ago,
the completion rate was 34 percent in the
humanities. Now it's 68 percent.
Attrition "destroys people's confidence
in themselves when they perceive themselves
as failures, when the problem should be
laid at other doors," says Mr. Thach.
"We don't want to be in the business
of disappointing people."
For Ms. Lovitts, tackling the problem of
attrition means that everyone involved --
from deans to department chairs to faculty
advisers -- must take more responsibility
for what happens to their graduate students.
"My personal feeling is that when a
university admits a graduate student to
Yet the pot of gold at the end of the Ph.D.
rainbow may not be there for every candidate.
For many of them, despite their love of
the subject and their dreams of reveling
in the life of the mind, the most logical
decision may be to leave.
After a year in a Ph.D. program in history
at City University of New York, Nicole Kalian
left to take a job as a publicist with a
book publisher. Hers was the sort of early
attrition that almost everyone agrees is
the best kind.
"I didn't see any prospects for when I graduated," says Ms. Kalian, who was shocked to read an article about new Ph.D.'s who couldn't find jobs as adjuncts on enough campuses to earn at least $25,000 a year. "It was frightening, and I could never really shake that thought from my head."
Copyright 2004 by The Chronicle of Higher Education
For problems or comments about page contact Richard Cherwitz