phd dropout rate

Who drops out of the Ph.D.?

Aside from the fact that doing a Ph.D. seems to put you at a greater risk of being anxious or depressed than other occupations, some students may also face the question: will I ever finish my thesis at all ? This post digs into research about doctoral attrition and completion, and what factors seem to make dropping out more likely. Do not give up!… unless you really want to.

Image by Siggy Nowak from Pixabay

About 10 years ago, when I left my job in the telecom industry to pursue an academic Ph.D., I started pursuing my other secret dream: being a psychologist. I enrolled in an online masters program on Psychology research. However, several months into the program, it was clear that this was not going to be a piece of cake. I was behind on the readings, some of the concepts in the courses were incomprehensible to me (not surprising, since my background was in Engineering), and I had some unpleasant online interactions with my peers when seeking help about these issues. My morale started to falter, and I started wondering: should I cut my losses and focus on my other main project (the doctoral degree that I had started in parallel)? Or was it better to drop out of the Ph.D. and pursue the shorter, maybe more manageable masters degree?

Dropping out of the PhD: the problem of attrition

If you ever faced this kind of thoughts, you are not alone. Not at all . In the research literature about students dropping out of doctoral programs (or “attrition”, as they call it), very often the ballpark of 40–60% attrition rate is mentioned 1 . Imagine you are in a classroom with your peer Ph.D. students. Look to the person on your left. Look to the person on your right. According to the statistics, only one or two of you will ever finish the Ph.D.

That’s a hard pill to swallow.

Of course, this is just a general approximation. The numbers vary quite a bit from university to university, and across the different disciplines: in one study, students in science and technology were 50% more likely to complete their Ph.D.’s than health sciences ones, and more than twice as likely compared to doctoral students in the humanities and social sciences 2 . This is probably due to many social, economic and cultural factors that are quite different in each discipline (rather than the inherent difficulty of the subject). Furthermore, even getting to these numbers is quite hard, since very often the researchers running these studies (and the doctoral programs themselves) don’t have a good way to know if a student has actually dropped out, or is just unusually quiet.

There is also the issue of when will you drop out. Several studies mention that dropping out of a Ph.D. is more probable in the first two years 1 , 2 . This is probably due to the students coming to the doctorate with a certain image or expectation of what doing research looks like, and academic research life not living up to those expectations – leading to disillusionment and dropout. To avoid this, in certain areas like biomedical research, students spend some time at the beginning of the doctorate rotating around different labs to get a better sense of what working in research looks like… only sometimes this also backfires, when labs start competing fiercely for the best students, so that some labs show a “friendly façade” during rotation, and a much harsher reality once the student incorporates to the lab for real 3 .

So, it is clear now: if about half of the doctoral students actually drop out of the Ph.D., probably many more have at least considered quitting. Indeed, I’d wager that you are quite lucky if you have not thought of abandoning the Ph.D. so far.

Factors for attrition

Now that we know this is a quite common problem, what are the factors related to greater chances of dropping out (or persisting until completion)? Looking into the research on this issue, I found five factors that appear quite often 4 :

• Kind of funding : Where you get the money from to do your Ph.D. is one of the most studied variables in relation to doctoral students dropping out. While the details vary from country to country and from one discipline to another (which determines what options are available to you), in general having no funding is associated with the highest rates of dropout (between 2.5 and six times more likely to drop out than the other options) 2 . Scholarships or research assistantships seem associated with lower dropouts 5 . While details differ across studies, it seems that the more your money source is aligned with your Ph.D., the better (e.g., if you have a scholarship that lets you freely choose your research topic, or your salary comes from a research project fully aligned with that topic). Conversely, if you are doing your Ph.D. part-time (because you took a completely unrelated job to pay your bills, or if you took a part-time lectureship at the university), you might be setting yourself up to have a harder time finishing your doctorate 6 .
• Marital status : Interestingly, married doctoral students (or those in a long-term relationship) are much less likely to drop out of a Ph.D. 5 . For example, in one study in Belgium, researchers found that married students were about four times less likely to drop out than unmarried ones 2 . As we saw when looking at depression and anxiety , it seems that this kind of close social (and, maybe, economic) support is very helpful to persist in the long journey of the doctorate.
• Career prospects . Quite logically, if you think that you will easily find a job once you have the doctorate under your belt, you will be more likely to persist until completion 5 . And vice-versa: once you stop believing that you have good career opportunities after finishing the Ph.D, dropping out seems much more likely 1 .
• Relationship with the supervisor . Rivers of ink have been poured trying to explain the influence of supervisors, and their relationship with students, in doctoral attrition. The issue appears in most qualitative accounts of doctoral student dropouts, but so far it has been quite difficult to quantify (e.g., well-advised students also drop out sometimes 3 ). Reviews of this area indicate that a positive student-supervisor relationship is associated with completing the Ph.D. 1 . But what makes a relationship positive? that’s harder to say… the advisors being “available”, having frequent interactions, having a sense of cooperation, understanding, and trust. Others also mention a non-hierarchical relationship, with clear expectations, or the advisor not being over-involved in their own research agenda 5 . Being an advisor myself, I find some of this advice wonderfully vague, but I will dig deeper into the topic of supervision styles, in a later post.
• Candidate “preparedness” is another factor that pops-up quite frequently, both in the sense of prior academic achievement (e.g., whether you passed your masters with very high grades), but also in terms of other personal characteristics of the doctoral student. For instance, in one study researchers found that students that passed the masters with very high distinction were two to eight times more likely to complete the Ph.D. than students that had their masters with lower grades 2 . However, not every review agrees that academic achievement is a critical factor in Ph.D. attrition 3 . Other reviews also mention students’ personal characteristics, such as their motivation to do the Ph.D. (if the motivation is high, and is about learning or personal improvement, chances of completion seem to be higher), students’ time on task, not having negative personal issues, etc. 5

OK, so far things seem logical. If we are alone, we are not academic over-achievers, or we have to get an unrelated job to make ends meet, we will generally have a harder time during the dissertation – and more chances of being faced with the dilemma of abandoning it.

But… what is the right answer?

A contrarian view of Ph.D. dropout

Most of the research I’ve seen around this topic describes dropping out as a big problem, a waste of time and resources for everyone involved (students, supervisors, universities, society). And, don’t get me wrong, I totally see how it is a problem that should not be dismissed lightly. However, I cannot help but think that we are seeing only one side of the coin: that of the institutional success, and the student as a human resource . We could also be a bit more empathetic and look at students as a human beings , and their experience: what if dropping out is the better option for this particular student, as a person, at this point in time? In one paper, a doctoral student explains:

‘‘I discussed withdrawing with family and my significant other; they just wanted me to be happy and, given the treatment that I received [from my advisor] for months, it seemed like the clear choice’’ 3

The quote reads like a really well thought-out, meditated decision, after enduring a toxic situation – regardless of the resources “wasted”.

Plus, are they really wasted? We may be forgetting that, even if you do not have a paper calling you “Ph.D.”, it is quite probable that you learned a few useful things during this journey, however incomplete: you learned to read scientific papers, you learned how your kind of research is really made, you learned to write and to argue a bit better, and probably you also practiced your critical thinking (which seems in short supply these days). I wouldn’t call that a total waste.

So yes, you should consider carefully before starting a Ph.D. (or accepting to supervise one). But, if the decision was made in good faith, forget about the funding, forget about the time “wasted”… they are sunk costs 7 . Rather, ask yourself: am I (or is this person) going to be an effective, convinced, purposeful researcher, if I continue my doctoral training under these conditions and in this place? If the answer is no, then maybe quitting isn’t a such bad idea. Heck, there is even research that suggests that, if you are at the point where you could decide by tossing a coin, you would be better off making the change right away! 8

If you are facing this conundrum, evaluate your environment and your daily experience carefully, and talk about it with family and close friends. But the decision is only yours. Yet, I can give you a general rule of thumb, from what I’ve seen in the academic world so far: if you think you are not “smart enough”, or you have any other argument for why you will never succeed at this that smells even remotely of impostor syndrome , I’d say you can make it (believe me, I’ve seen some really un-smart people get doctorates). If, on the other hand, your lab environment is toxic, your economic or social situation is really bad, or you feel deeply unhappy every day you do research, maybe it is time for a re-evaluation.

You can do it, if you want to endure (or -gasp!- enjoy) the process.

Coming back to my own personal case, I did drop out of the Psychology masters, to focus on my Ph.D. And I don’t regret it one bit. Indeed, even after focusing my attention on the Ph.D., a researcher could have told me that my chances were still not terribly optimistic: I was single, I was completely self-funded, my masters grades were not exactly glowing, and I had no idea whether the doctorate would bring me incredible job opportunities.

Oddly enough, not only I managed to finish my Ph.D.; I actually consider that year one of the happiest, most fulfilling of my life.

Am I an outlier? Maybe yes. Was I extremely lucky? Probably so. However, in some of the latest readings I did for this post, I found an alternative, reasonable explanation. But this post has gotten quite long already. You can find out more about this other strand of doctoral education research, in the next post of the series on doctoral dropout .

Have you ever considered dropping out of your Ph.D.? can you think of other factors that made you stay (or abandon it)? Do you think there is a right moment to quit the doctorate? Let me know in the comments section below!

New to the blog? Read more about…

Common problems and challenges in doing the PhD, from mental health (e.g., depression or anxiety) or productivity challenges , to writing or dropping out of your PhD .

Mental health and wellbeing tips and advice : common mental health symptoms in the PhD , tips to avoid dropping out of the doctorate , the importance of sleep , holidays or advice from positive psychology to keep yourself motivated during the PhD.

PhD productivity tips and advice : from the classic Pomodoro technique , to avoiding to-do list overwhelm , dealing with multiple projects and priorities , staying productive and motivated , how I manage my daily tasks or how I do my weekly reviews .

PhD-specific tools , like the CQOCE diagram to conceptualize your PhD, the NABC method to structure your research presentations, or the process I use to write scientific papers or make big career decisions .

Supervision tips and advice , about giving feedback on student papers , or supporting a sense of progress in your doctoral students .

See, for example, Bair, C. R., & Haworth, J. G. (2004). Doctoral student attrition and persistence: A meta-synthesis of research. In Higher education: Handbook of theory and research (pp. 481–534). Springer. ↩︎

Wollast, R., Boudrenghien, G., Van der Linden, N., Galand, B., Roland, N., Devos, C., … Frenay, M. (2018). Who Are the Doctoral Students Who Drop Out? Factors Associated with the Rate of Doctoral Degree Completion in Universities. International Journal of Higher Education , 7 (4), 143–156. ↩︎

Maher, M. A., Wofford, A. M., Roksa, J., & Feldon, D. F. (2017). Exploring Early Exits: Doctoral Attrition in the Biomedical Sciences. Journal of College Student Retention: Research, Theory & Practice . https://doi.org/10.1177/1521025117736871 ↩︎

Please be aware that most of this evidence is from correlational studies, so it is hard to know if these factors are the causes of the dropout, or (more probably) symptoms of a different underlying cause (or causes). ↩︎

Rigler Jr, K. L., Bowlin, L. K., Sweat, K., Watts, S., & Throne, R. (2017). Agency, Socialization, and Support: A Critical Review of Doctoral Student Attrition. Paper Presented at the 3rd International Conference on Doctoral Education . Presented at the University of Central Florida. Retrieved from https://files.eric.ed.gov/fulltext/ED580853.pdf ↩︎

Gardner, S. K., & Gopaul, B. (2012). The part-time doctoral student experience. International Journal of Doctoral Studies , 7 (12), 63–78. Retrieved from http://informingscience.com/ijds/Volume7/IJDSv7p063-078Gardner352.pdf ↩︎

Arkes, H. R., & Blumer, C. (1985). The psychology of sunk cost. Organizational Behavior and Human Decision Processes , 35 (1), 124–140. ↩︎

Levitt, S. D. (2016). Heads or tails: The impact of a coin toss on major life decisions and subsequent happiness (Working Paper No. 22487). Retrieved from National Bureau of Economic Research website: https://www.nber.org/papers/w22487 ↩︎

Luis P. Prieto

Luis P. is a Ramón y Cajal research fellow at the University of Valladolid (Spain), investigating learning technologies, especially learning analytics. He is also an avid learner about doctoral education and supervision, and he's the main author at the A Happy PhD blog.

Google Scholar profile

• PhD Failure Rate – A Study of 26,076 PhD Candidates
• Doing a PhD

The PhD failure rate in the UK is 19.5%, with 16.2% of students leaving their PhD programme early, and 3.3% of students failing their viva. 80.5% of all students who enrol onto a PhD programme successfully complete it and are awarded a doctorate.

Introduction

One of the biggest concerns for doctoral students is the ongoing fear of failing their PhD.

After all those years of research, the long days in the lab and the endless nights in the library, it’s no surprise to find many agonising over the possibility of it all being for nothing. While this fear will always exist, it would help you to know how likely failure is, and what you can do to increase your chances of success.

Read on to learn how PhDs can be failed, what the true failure rates are based on an analysis of 26,067 PhD candidates from 14 UK universities, and what your options are if you’re unsuccessful in obtaining your PhD.

Ways You Can Fail A PhD

There are essentially two ways in which you can fail a PhD; non-completion or failing your viva (also known as your thesis defence ).

Non-completion

Non-completion is when a student leaves their PhD programme before having sat their viva examination. Since vivas take place at the end of the PhD journey, typically between the 3rd and 4th year for most full-time programmes, most failed PhDs fall within the ‘non-completion’ category because of the long duration it covers.

There are many reasons why a student may decide to leave a programme early, though these can usually be grouped into two categories:

• Motives – The individual may no longer believe undertaking a PhD is for them. This might be because it isn’t what they had imagined, or they’ve decided on an alternative path.
• Extenuating circumstances – The student may face unforeseen problems beyond their control, such as poor health, bereavement or family difficulties, preventing them from completing their research.

In both cases, a good supervisor will always try their best to help the student continue with their studies. In the former case, this may mean considering alternative research questions or, in the latter case, encouraging you to seek academic support from the university through one of their student care policies.

Besides the student deciding to end their programme early, the university can also make this decision. On these occasions, the student’s supervisor may not believe they’ve made enough progress for the time they’ve been on the project. If the problem can’t be corrected, the supervisor may ask the university to remove the student from the programme.

Failing The Viva

Assuming you make it to the end of your programme, there are still two ways you can be unsuccessful.

The first is an unsatisfactory thesis. For whatever reason, your thesis may be deemed not good enough, lacking originality, reliable data, conclusive findings, or be of poor overall quality. In such cases, your examiners may request an extensive rework of your thesis before agreeing to perform your viva examination. Although this will rarely be the case, it is possible that you may exceed the permissible length of programme registration and if you don’t have valid grounds for an extension, you may not have enough time to be able to sit your viva.

The more common scenario, while still being uncommon itself, is that you sit and fail your viva examination. The examiners may decide that your research project is severely flawed, to the point where it can’t possibly be remedied even with major revisions. This could happen for reasons such as basing your study on an incorrect fundamental assumption; this should not happen however if there is a proper supervisory support system in place.

PhD Failure Rate – UK & EU Statistics

According to 2010-11 data published by the Higher Education Funding Council for England (now replaced by UK Research and Innovation ), 72.9% of students enrolled in a PhD programme in the UK or EU complete their degree within seven years. Following this, 80.5% of PhD students complete their degree within 25 years.

This means that four out of every five students who register onto a PhD programme successfully complete their doctorate.

While a failure rate of one in five students may seem a little high, most of these are those who exit their programme early as opposed to those who fail at the viva stage.

Failing Doesn’t Happen Often

Although a PhD is an independent project, you will be appointed a supervisor to support you. Each university will have its own system for how your supervisor is to support you , but regardless of this, they will all require regular communication between the two of you. This could be in the form of annual reviews, quarterly interim reviews or regular meetings. The majority of students also have a secondary academic supervisor (and in some cases a thesis committee of supervisors); the role of these can vary from having a hands-on role in regular supervision, to being another useful person to bounce ideas off of.

These frequent check-ins are designed to help you stay on track with your project. For example, if any issues are identified, you and your supervisor can discuss how to rectify them in order to refocus your research. This reduces the likelihood of a problem going undetected for several years, only for it to be unearthed after it’s too late to address.

In addition, the thesis you submit to your examiners will likely be your third or fourth iteration, with your supervisor having critiqued each earlier version. As a result, your thesis will typically only be submitted to the examiners after your supervisor approves it; many UK universities require a formal, signed document to be submitted by the primary academic supervisor at the same time as the student submits the thesis, confirming that he or she has approved the submission.

Failed Viva – Outcomes of 26,076 Students

Despite what you may have heard, the failing PhD rate amongst students who sit their viva is low.

This, combined with ongoing guidance from your supervisor, is because vivas don’t have a strict pass/fail outcome. You can find a detailed breakdown of all viva outcomes in our viva guide, but to summarise – the most common outcome will be for you to revise your thesis in accordance with the comments from your examiners and resubmit it.

This means that as long as the review of your thesis and your viva examination uncovers no significant issues, you’re almost certain to be awarded a provisional pass on the basis you make the necessary corrections to your thesis.

To give you an indication of the viva failure rate, we’ve analysed the outcomes of 26,076 PhD candidates from 14 UK universities who sat a viva between 2006 and 2017.

The analysis shows that of the 26,076 students who sat their viva, 25,063 succeeded; this is just over 96% of the total students as shown in the chart below.

Students Who Passed

The analysis shows that of the 96% of students who passed, approximately 5% required no amendments, 79% required minor amendments and the remaining 16% required major revisions. This supports our earlier discussion on how the most common outcome of a viva is a ‘pass with minor amendments’.

Students Who Failed

Of the 4% of unsuccessful students, approximately 97% were awarded an MPhil (Master of Philosophy), and 3% weren’t awarded a degree.

Note : It should be noted that while the data provides the student’s overall outcome, i.e. whether they passed or failed, they didn’t all provide the students specific outcome, i.e. whether they had to make amendments, or with a failure, whether they were awarded an MPhil. Therefore, while the breakdowns represent the current known data, the exact breakdown may differ.

Summary of Findings

By using our data in combination with the earlier statistic provided by HEFCE, we can gain an overall picture of the PhD journey as summarised in the image below.

To summarise, based on the analysis of 26,076 PhD candidates at 14 universities between 2006 and 2017, the PhD pass rate in the UK is 80.5%. Of the 19.5% of students who fail, 3.3% is attributed to students failing their viva and the remaining 16.2% is attributed to students leaving their programme early.

The above statistics indicate that while 1 in every 5 students fail their PhD, the failure rate for the viva process itself is low. Specifically, only 4% of all students who sit their viva fail; in other words, 96% of the students pass it.

What Are Your Options After an Unsuccessful PhD?

Appeal your outcome.

If you believe you had a valid case, you can try to appeal against your outcome . The appeal process will be different for each university, so ensure you consult the guidelines published by your university before taking any action.

While making an appeal may be an option, it should only be considered if you genuinely believe you have a legitimate case. Most examiners have a lot of experience in assessing PhD candidates and follow strict guidelines when making their decisions. Therefore, your claim for appeal will need to be strong if it is to stand up in front of committee members in the adjudication process.

Downgrade to MPhil

If you are unsuccessful in being awarded a PhD, an MPhil may be awarded instead. For this to happen, your work would need to be considered worthy of an MPhil, as although it is a Master’s degree, it is still an advanced postgraduate research degree.

Unfortunately, there’s a lot of stigma around MPhil degrees, with many worrying that it will be seen as a sign of a failed PhD. While not as advanced as a PhD, an MPhil is still an advanced research degree, and being awarded one shows that you’ve successfully carried out an independent research project which is an undertaking to be admired.

Finding a PhD has never been this easy – search for a PhD by keyword, location or academic area of interest.

Additional Resources

Hopefully now knowing the overall picture your mind will feel slightly more at ease. Regardless, there are several good practices you can adopt to ensure you’re always in the best possible position. The key of these includes developing a good working relationship with your supervisor, working to a project schedule, having your thesis checked by several other academics aside from your supervisor, and thoroughly preparing for your viva examination.

We’ve developed a number of resources which should help you in the above:

• What to Expect from Your Supervisor – Find out what to look for in a Supervisor, how they will typically support you, and how often you should meet with them.
• How to Write a Research Proposal – Find an outline of how you can go about putting a project plan together.
• What is a PhD Viva? – Learn exactly what a viva is, their purpose and what you can expect on the day. We’ve also provided a full breakdown of all the possible outcomes of a viva and tips to help you prepare for your own.

Data for Statistics

• Cardiff University – 2006/07 to 2016/17
• Imperial College London – 2006/07 to 2016/17
• London School of Economics (LSE) – 2006/07 to 2015/16
• Queen Mary University of London – 2009/10 to 2015/16
• University College London (UCL) – 2006/07 to 2016/17
• University of Aberdeen – 2006/07 to 2016/17
• University of Birmingham – 2006/07 to 2015/16
• University of Bristol – 2006/07 to 2016/17
• University of Edinburgh – 2006/07 to 2016/17
• University of Nottingham – 2006/07 to 2015/16
• University of Oxford – 2007/08 to 2016/17
• University of York – 2009/10 to 2016/17
• University of Manchester – 2008/09 to 2017/18
• University of Sheffield – 2006/07 to 2016/17

Note : The data used for this analysis was obtained from the above universities under the Freedom of Information Act. As per the Act, the information was provided in such a way that no specific individual can be identified from the data.

Browse PhDs Now

Join thousands of students.

Join thousands of other students and stay up to date with the latest PhD programmes, funding opportunities and advice.

National Center for Science and Engineering Statistics

• All previous cycle years

The SED is an annual census of research doctorate recipients from U.S. academic institutions that collects information on educational history, demographic characteristics, graduate funding source and educational debts, and postgraduation plans.

Survey Info

• tag for use when URL is provided --> Methodology
• tag for use when URL is provided --> Data
• tag for use when URL is provided --> Analysis

The Survey of Earned Doctorates is an annual census conducted since 1957 of all individuals receiving a research doctorate from an accredited U.S. institution in a given academic year. The SED is sponsored by the National Center for Science and Engineering Statistics (NCSES) within the National Science Foundation (NSF) and by three other federal agencies: the National Institutes of Health, Department of Education, and National Endowment for the Humanities. The SED collects information on the doctoral recipient’s educational history, demographic characteristics, and postgraduation plans. Results are used to assess characteristics of the doctoral population and trends in doctoral education and degrees.

Areas of Interest

• STEM Education
• Innovation and Global Competitiveness

Survey Administration

The 2022 survey was conducted by RTI International under contract to NCSES.

Survey Details

Featured survey analysis.

Doctorate Recipients from U.S. Universities: 2022

SED Overview

Data highlights, the number of research doctorates conferred by u.s. institutions, which began a sharp 15-month decline in spring 2020 due to the covid-19 pandemic, rebounded in 2022 with the highest number of research doctorates awarded in any academic year to date.

Over the past 20 years, most of the growth in the number of doctorates earned by both men and women has been in science and engineering (S&E) fields 

Methodology

Survey description, technical notes, technical tables, questionnaires, view archived questionnaires, featured analysis.

Research Doctorate Conferrals Rebound, Leading to Record Number of U.S. Doctorate Recipients in 2022

Related content, related collections, survey contact.

For additional information about this survey or the methodology, contact

Get e-mail updates from NCSES

NCSES is an official statistical agency. Subscribe below to receive our latest news and announcements.

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings

Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .

• Advanced Search
• Journal List
• HHS Author Manuscripts

EXPLORING REASONS THAT U.S. MD-PHD STUDENTS ENTER AND LEAVE THEIR DUAL-DEGREE PROGRAMS

Devasmita chakraverty.

Indian Institute of Management Ahmedabad, Ahmedabad, India

Donna B. Jeffe

Washington University in St. Louis, St. Louis, MO, U.S.A.

Katherine P. Dabney

Virginia Commonwealth University, Richmond, VA, U.S.A.

Robert H. Tai

University of Virginia, Charlottesville, VA, U.S.A.

Aim/Purpose

In response to widespread efforts to increase the size and diversity of the biomedical-research workforce in the U.S., a large-scale qualitative study was conducted to examine current and former students’ training experiences in MD (Doctor of Medicine), PhD (Doctor of Philosophy), and MD-PhD dual-degree programs. In this paper, we aimed to describe the experiences of a subset of study participants who had dropped out their MD-PhD dual-degree training program, the reasons they entered the MD-PhD program, as well as their reasons for discontinuing their training for the MD-PhD.

The U.S. has the longest history of MD-PhD dual-degree training programs and produces the largest number of MD-PhD graduates in the world. In the U.S., dual-degree MD-PhD programs are offered at many medical schools and historically have included three phases—preclinical, PhD-research, and clinical training, all during medical-school training. On average, it takes eight years of training to complete requirements for the MD-PhD dual-degree. MD-PhD students have unique training experiences, different from MD-only or PhD-only students. Not all MD-PhD students complete their training, at a cost to funding agencies, schools, and students themselves.

Methodology

We purposefully sampled from 97 U.S. schools with doctoral programs, posting advertisements for recruitment of participants who were engaged in or had completed PhD, MD, and MD-PhD training. Between 2011-2013, semi-structured, one-on-one phone interviews were conducted with 217 participants. Using a phenomenological approach and inductive, thematic analysis, we examined students’ reasons for entering the MD-PhD dual-degree program, when they decided to leave, and their reasons for leaving MD-PhD training.

Contribution

Study findings offer new insights into MD-PhD students’ reasons for leaving the program, beyond what is known about program attrition based on retrospective analysis of existing national data, as little is known about students’ actual reasons for attrition. By more deeply exploring students’ reasons for attrition, programs can find ways to improve MD-PhD students’ training experiences and boost their retention in these dual-degree programs to completion, which will, in turn, foster expansion of the biomedical-research-workforce capacity.

Seven participants in the larger study reported during their interview that they left their MD-PhD programs before finishing, and these were the only participants who reported leaving their doctoral training. At the time of interview, two participants had completed the MD and were academic-medicine faculty, four were completing medical school, and one dropped out of medicine to complete a PhD in Education. Participants reported enrolling in MD-PhD programs to work in both clinical practice and research. Very positive college research experiences, mentorship, and personal reasons also played important roles in participants’ decisions to pursue the dual MD-PhD degree. However, once in the program, positive mentorship and other opportunities that they experienced during or after college, which initially drew candidates to the program was found lacking. Four themes emerged as reasons for leaving the MD-PhD program: 1) declining interest in research, 2) isolation and lack of social integration during the different training phases, 3) suboptimal PhD-advising experiences, and 4) unforeseen obstacles to completing PhD research requirements, such as loss of funding.

Recommendations for Practitioners

Though limited by a small sample size, findings highlight the need for better integrated institutional and programmatic supports for MD-PhD students, especially during PhD training.

Recommendations for Researchers

Researchers should continue to explore if other programmatic aspects of MD-PhD training (other than challenges experienced during PhD training, as discussed in this paper) are particularly problematic and pose challenges to the successful completion of the program.

Impact on Society

The MD-PhD workforce comprises a small, but highly -trained cadre of physician-scientists with the expertise to conduct clinical and/or basic science research aimed at improving patient care and developing new diagnostic tools and therapies. Although MD-PhD graduates comprise a small proportion of all MD graduates in the U.S. and globally, about half of all MD-trained physician-scientists in the U.S. federally funded biomedical-research workforce are MD-PhD-trained physicians. Training is extensive and rigorous. Improving experiences during the PhD-training phase could help reduce MD-PhD program attrition, as attrition results in substantial financial cost to federal and private funding agencies and to medical schools that fund MD-PhD programs in the U.S. and other countries.

Future Research

Future research could examine, in greater depth, how communications among students, faculty and administrators in various settings, such as classrooms, research labs, and clinics, might help MD-PhD students become more fully integrated into each new program phase and continue in the program to completion. Future research could also examine experiences of MD-PhD students from groups underrepresented in medicine and the biomedical-research workforce (e.g., first-generation college graduates, women, and racial/ethnic minorities), which might serve to inform interventions to increase the numbers of applicants to MD-PhD programs and help reverse the steady decline in the physician-scientist workforce over the past several decades.

Introduction

Traditional doctoral training for the PhD involves time for trainees to learn to combine their knowledge of course content and research skills to produce original research, culminating with a doctoral dissertation ( Lovitts, 2005 ). Typically, the average time of PhD-degree completion varies from 4-6 years ( Bourke et al., 2004 ). The MD-PhD (Doctor of Medicine and Doctor of Philosophy) physician-scientist workforce comprises a relatively small cadre of well-trained physician-scientists with the research skills to address clinical and/or basic science research questions aimed at improving patient care ( Goldstein & Kohrt, 2012 ; Varki & Rosenberg, 2002 ). In the U.S., MD-PhD training during medical school is extensive and lengthy, typically lasting for eight or more years ( Brass et al., 2010 ; Jeffe et al., 2014a ), and MD-PhD program attrition is a cause of concern. To our knowledge, only one study has been conducted to examine factors associated with MD-PhD program attrition ( Jeffe et al., 2014a ), and no studies have purposely examined MD-PhD students’ own reasons for leaving their MD-PhD program.

To fill a gap in the literature, we examined attrition from MD-PhD training programs in the U.S., where such training programs were first developed in the 1950s to increase the number of physician-scientists in the biomedical-research workforce (Harding et al., 2017) and where integrated dual-degree MD-PhD programs are the most prevalent. For the award period from July 1, 2019 through June 30, 2020, 50 of 154 U.S. Liaison Committee on Medical Education (LCME)-accredited medical schools had dual-degree MD-PhD programs that were funded by the U.S. National Institutes of Health National Institute of General Medical Sciences (NIH NIGMS) Medical Scientist Training Program (MSTP) ( National Institute of General Medical Sciences, 2020 ). Many, if not all, MSTP-funded MD-PhD programs as well as non-MSTP-funded MD-PhD programs in U.S. medical schools receive training support from non-federal governmental and private funding organizations, other NIH institutes, and institutional funds to support MD-PhD training ( AAMC, 2009 ; Jeffe et al., 2014a ; Jeffe & Andriole, 2011 ). MD-PhD programs in other countries are small in number relative to the number of MD-PhD programs in the U.S. ( Jones et al., 2016 ; Kuehnle et al., 2009 ; Twa et al., on behalf of the Canadian MD/PhD Program Investigation Group, 2017 ), and many of the nationally supported MD-PhD programs in other countries, such as Switzerland ( Kuehnle et al, 2009 ) and Germany ( Bossé et al., 2011 ), allow for PhD training to begin after receipt of the MD. A 2016-2017 survey of the European MD/PhD Association programs in multiple countries examined MD-PhD program characteristics in association with MD-PhD students’ and graduates’ opinions about the program, their career choices and outcomes ( dos Santos Rocha et al., 2020 ); but we found no studies published that examined MD-PhD students’ self-reported reasons for leaving the MD-PhD program prior to completion.

This exploratory study therefore sought to answer the following research questions: “For MD-PhD students who discontinued their training, what motivated them to pursue MD-PhD training? Additionally, at what point during training and for what reasons did they discontinue their training?”

Literature Review

Md-phd programs typically involve three phases:.

two years of pre-clinical training in medical school, at least four years of PhD research training in graduate school, and two more years of clinical training after returning to medical school ( Brass et al., 2010 ; Jeffe et al., 2014a ). Acceptance to MD-PhD dual-degree programs is very competitive, and MD-PhD graduates have a greater planned career involvement in research at the time of medical-school graduation compared with all other MD graduates ( Andriole et al., 2008 ), especially in disease-oriented and clinical research ( Ahn et al., 2007 ; Andriole et al., 2008 ).

Not all students who matriculate into MD-PhD programs complete the program ( Jeffe et al., 2014a ; National Institutes of Health National Institute of General Medical Sciences [NIH-NIGMS], 1998 ). In an earlier survey study, more than one-fourth of enrolled MD-PhD students seriously considered leaving the program ( Ahn et al., 2007 ). In a survey of 24 MD-PhD programs ( Brass et al., 2010 ), attrition rates were reported to range from 3-34%. In a national cohort study of MD-PhD program enrollees at time of matriculation, the attrition rate was observed to be 27% ( Jeffe et al., 2014a ). By comparison, the attrition rate among MD-only students in the U.S. is about 3% ( Association of American Medical Colleges [AAMC], 2012 ; Garrison et al., 2007 ). PhD enrollment and completion rates vary across universities, fields, countries, and demographic factors such as sex ( Dabney et al., Tai, 2016 ); national-level data on PhD-program attrition is not well-documented. An Australian study collected data from approximately 1,200 students enrolled at one university to find a completion rate of 70% ( Bourke et al., 2004 ). In another study, attrition data were collected in 2013-2014 in a survey of more than 1,500 psychology programs in the U.S. and found doctoral attrition rates between 5-13% ( Michalski et al., 2016 ). Dropout rates during PhD training have been reported to be between 40% and 60% ( Geiger, 1997 ; Tinto, 1987 ). The odds of PhD student dropout in STEM is most in the first year and greater for women ( Lott et al., 2009 ). One study about underrepresented racial/ethnic minority (URM, including Black/African American, Hispanic/Latino, and American Indian/Native American) students in STEM that collated PhD completion rates for ten years found Black and Hispanic students to have PhD completion rates of 50% and 58%, respectively ( Okahana et al., 2016 ).

While navigating the preclinical, research, and clinical phases of training, MD-PhD students face unique challenges different from MD-only or PhD-only students ( Chakraverty et al., 2018 ). More MD-PhD than MD students anticipate or experience challenges to balancing training and family life ( Kwan et al., 2017 ). Students also find that the tripartite model of MD-PhD dual-degree programs in the U.S. and Canada creates challenges, having to navigate two transitions between training phases ( Bossé et al., 2011 ; Chakraverty et al., 2018 ), which most students in MD-only or PhD-only programs do not experience. Among the challenges experienced by MD-PhD students having to transition between the phases are time away from the clinical environment, which could impact students’ preparedness for clinical clerkships ( Goldberg & Insel, 2013 ) as well as a lack of desired mentoring (especially mentoring by MD-PhD faculty), a perceived lack of curricular integration and of awareness of phase-specific cultural differences, and difficulties assimilating with other trainees during the research- and clinical-training phases, who are not from their original cohort of peers ( Chakraverty et al., 2018 ).

Large national cohort studies have examined educational experiences of MD-PhD students as well as variables associated with MD-PhD enrollment ( Jeffe et al., 2014b ), attrition ( Jeffe et al., 2014a ), and graduation ( Andriole et al., 2008 ). Individuals who reported participating in high school and college laboratory research apprenticeships, and who highly valued research and finding disease cures as the most important reason to study medicine were more likely to enroll in MD-PhD programs, demonstrating alignment of students’ attitudes and interests with MD-PhD program goals ( Jeffe et al., 2014b ; Tai et al., 2017 ). Students who planned substantial career involvement in research at graduation were more likely to be MD-PhD program graduates than all other-MD program graduates; controlling for other variables in the regression model, women and URM students were less likely to graduate from MD-PhD (vs. other-MD) programs ( Andriole et al., 2008 ). In another study of 2,582 MD-PhD program enrollees, 1,885 (73%) had completed the MD-PhD program, 597 (23%) dropped out of the program but completed the MD, and 100 (4%) left medical school entirely ( Jeffe et al., 2014a ). Although students who enrolled in MD-PhD programs at medical-school matriculation and planned substantial career involvement in research at that time were less likely to leave the MD-PhD program, students who had lower Medical College Admission Test scores, attended medical schools without NIH NIGMS MSTP-funded MD-PhD programs, and were older at matriculation were more likely to leave their MD-PhD program. Notably, women and URM students were neither more nor less likely to leave the MD-PhD program and graduate with only an MD degree ( Jeffe et al., 2014a ). Students’ MD-PhD program satisfaction was reported to be higher at the beginning of the program and lower during the research phase, due to the unpredictability of time to complete the PhD ( Ahn et al., 2007 ).

Although research has examined challenges faced by potential MD-PhD program applicants ( Kersbergen et al., 2020 ) and by MD-PhD students during their training as described above, to our knowledge, no study has examined the reasons why MD-PhD students leave the program before completing their training using a qualitative research approach. Qualitative research can help explain the decision-making process of individuals ( Marshall & Rossman, 2006 ), adding to our understanding of reasons for leaving the program from participants’ perspectives of their personal experiences. We examined attrition from MD-PhD dual-degree programs using a lens of integration and interaction ( Kong et al., 2013 ) to better understand why some U.S. MD-PhD students ultimately discontinued their training.

The data for this paper were collected for a larger qualitative study (Transitions in the Education of Minorities Underrepresented in Research) conducted in the U.S. between 2010 and 2014. This larger study examined training experiences of doctoral students and postdoctoral trainees planning to pursue careers in the biomedical-research workforce to identify factors that served to facilitate or impede progress along this career path ( Andriole et al., 2015 ; Chakraverty, 2013 ; Chakraverty et al., 2018 ; Jeffe et al., 2014a ; Jeffe et al., 2014b ; Tai et al., 2017 ). In all, we conducted 217 interviews with PhD, MD, and MD-PhD students, postdocs, physician-scientists, and faculty in U.S. higher education biomedical-science PhD programs and in MD-PhD dual-degree programs in U.S. medical schools.

Methodological considerations for conducting a qualitative study were governed by the aims of the larger study to more deeply understand participants’ reasons for considering doctoral-level training in the biomedical sciences in pursuit of a research career and for attrition from MD-PhD training specifically, if applicable, which is the focus of the current study. Using a phenomenological approach, we examined how participants made their decisions to enter or leave their training programs ( Marshall & Rossman, 2006 ). Semi-structured, in-depth interviews allowed us to gather detailed narratives to learn more about all participants’ decision-making processes to enter and either complete or leave their doctoral training ( DiCicco-Bloom & Crabtree, 2006 ). Although this paper focuses on attrition from the MD-PhD program, we also analyzed data for these participants’ reasons for enrolling in the MD-PhD program, to gain a more holistic understanding of their experiences and decision-making processes.

Data Collection and Analysis

Study sample and eligibility.

Following Institutional Review Board approval at the University of Virginia and Washington University in St. Louis, we purposefully sampled ( Marshall & Rossman, 2006 ; Miles & Huberman, 1994 ) U.S. public and private higher education institutions offering biomedical-science PhD degrees and medical schools with dual-degree MD-PhD programs. We sought to interview individuals training for or currently engaged in biomedical research; we also wanted to interview MD-PhD program trainees who dropped out of their program before graduation. We included higher education institutions with the Carnegie classification ( The Carnegie Classification of Institutions of Higher Education, n.d. ), indicating high or very high research activity. Deans and department chairs disseminated information about the study with our contact information, using emails, announcements, posters, and flyers. We also recruited participants through snowball sampling (Bogdan & Biklen, 2007; Sadler et al., 2010 ), asking current participants if they would be willing to share our contact information with their colleagues or other students in their program, as well as with individuals who had left their program, and to encourage them to participate in this study. We scheduled phone interviews with individuals who contacted us expressing an interest to participate.

Of 217 participants interviewed in the larger study, 29 students were then currently enrolled in an MD-only program, 20 in a PhD-only program, and 68 in an MD-PhD program; in addition, 25 participants were postdoctoral trainees at the time of the study. Participants no longer in school included 56 faculty, 14 non-scientists, 4 scientists outside academia, and one participant who dropped out of the MD-PhD program before completing either degree. For the current study about MD-PhD program attrition, anyone who had once enrolled in an MD-PhD program but did not complete it was eligible to participate. Overall, seven participants had been enrolled in dual-degree MD-PhD programs but subsequently discontinued MD-PhD training, six of whom continued their training for the MD. The current analysis examines the training experiences of those seven participants and reasons for discontinuing MD-PhD training, which was a specific aim of the larger study.

Semi-structured interviews

A semi-structured interview format allowed us some flexibility in asking question better tailored to an individual’s life experiences ( Cohen & Crabtree, 2006 ), although we asked everyone a basic set of questions, ( Table 1 ). Each participant completed one, 45-60 minute semi-structured telephone interview following their informed consent. The interview questions were developed based on the overall study aims, one of which focused on reasons for MD-PhD attrition. The interview protocol and questions were developed by the principal investigators and co-investigators based on their knowledge of gaps in the literature and understanding of the field; interview questions were reviewed by content experts and pilot tested before the initiation of data collection.

Interview Questions Asked of Participants

Specially trained interviewers, including faculty and PhD students on the research team, conducted interviews for this study. Demographic data such as age, sex, race/ethnicity, and current program were collected at the beginning of each interview. Interviews were audio-recorded with permission, transcribed verbatim through a professional company, and assigned an alpha-numeric code prior to analysis. For this aim of the study focusing on MD-PhD students’ reasons for leaving the MD-PhD program, in-depth interviews were conducted to gain insight into participants’ backgrounds, experiences, reasons for enrolling in MD-PhD programs, and when and why they discontinued their MD-PhD training through their own narratives ( Kvale & Brinkmann, 2009 ). We sought to identify aspects of the training that might have been particularly problematic for these participants. Probing questions were asked based on participants’ responses. All the authors have directly conducted or aided in medical education research for varying lengths of time. At the end of each interview, participants were asked to broadly share information about the study in their professional and personal networks, so that people from a wider network would become aware of this study using snowball sampling (Bogdan & Biklen, 2007; Sadler et al., 2010 ).

Analytic strategy

Each interview transcript was open-coded by two authors, both for narratives about their reasons for enrolling in the MD-PhD program and for leaving the program. The coders created a single codebook after discussing and resolving disagreements about codes, compiling all the codes into a final list that was used to reanalyze all the interviews. Since attrition MD-PhD program attrition is a relatively understudied topic, codes were based on participant transcripts rather than existing literature. Using an inductive, thematic approach as the primary analysis strategy ( Miles & Huberman, 1994 ; Pope et al., 2000 ) and the constant comparative method of coding ( Glaser & Strauss, 1967 ), the codes were systematically organized into themes ( Thomas, 2006 ). Themes that emerged from the analysis are presented if experiences fitting in a theme were discussed by multiple participants. Although some reasons described during the interview were unique for a participant, we elaborate only on those recurrent themes and experiences that were common across multiple participants. Both coders were mindful of the fact that their worldviews and positionalities could differ from those of the participants, interviewers, and from each other, which could influence how the interviews were conducted and data were analyzed ( Antin et al., 2015 ). Both coders were a part of the interview team and are educational researchers with a background in higher education and medical education research; they used a reflective journal, recording memos to document their coding decisions during analysis and acknowledge any disconfirming evidence. The coders also consulted with each other to ensure agreement on coding. They resolved coding disagreements through a discussion and consensus. The coding and analysis process lasted roughly seven months. We present representative quotes that exemplified the emergent themes, adding content in brackets to clarify a participant’s narrative. We used pseudonyms for those participants whose results are described in this manuscript.

Of the seven participants who had left the MD-PhD program before completion, two had completed their MD training and held academic-medicine faculty positions at the time of their interview; four were still in medical school completing the MD degree, and one was completing a PhD in Education ( Table 2 ). Since the sample size was small, our findings are exploratory; we did not expect to reach data saturation, a stage when no new themes emerge as a result of further data collection ( Faulkner & Trotter, 2017 ; Glaser & Strauss, 1967 ). Although there were similarities in the reasons that participants gave for entering MD-PhD training, each participant described slightly different circumstances and stages during which they left MD-PhD training.

Participant Demographics, Timing of MD-PhD-program Attrition, and Status at Time of Interview

Why participants entered MD-PhD training?

We asked participants what inspired them to pursue MD-PhD training in the first place. All seven participants provided reasons that included both a desire to help people on a day-to-day basis through clinical practice and to more deeply engage in research. Having the MD-PhD dual degree was perceived as a way to broaden research opportunities to participate in clinical and other types of research as well as get access to patient populations. For all participants, the desire to pursue a research career grew from undergraduate research opportunities that they had experienced; such opportunities led to publishing and presenting at conferences, networking with established researchers, and getting to know “what their careers were like” (Debbie). Ben had “a pretty thorough research experience” in college where he “worked every summer in the research lab” and had already published research by the time he finished college. Aaron described an undergraduate mentor who was “a very good chemist and a wonderful teacher” who taught him “how research is done and the rewards of doing research.” A fulfilling college research experience also provided participants with the skills to handle research responsibilities, independently decide what experiments to conduct, and develop ownership of the work—factors that made participants consider studying for an MD-PhD.

In college, it became much more concrete, this idea that I wanted to do research and medicine, and try and incorporate the two. The experience gave me this little niche to be working in and got me really excited about what scientists do. (Eva)

During college, participants reported having opportunities to give presentations at national conferences and to gain insight into clinical experiences by shadowing physicians, and volunteering to help children with special needs. Such experiences shaped one’s desire to pursue MD-PhD as opposed to MD-only or PhD-only. Eva, who wanted to combine medical training with research training shared, “as much as I like the research and thinking about science, I wasn’t cut out to just be in the lab all the time by myself.”

Participants were also influenced by undergraduate mentors who provided hands-on research experience by “letting me have my own little section of the project.… He said, ‘Here's a part of the project. I want you to figure this thing out.’ I think that’s what really sparked my enthusiasm for basic science” (Francesca). Overall, Eva realized that receiving both the MD and PhD would help “produce new knowledge and provide independence” and the “thrill of discovery.” College mentors also helped select and apply to MD-PhD programs and provided information about how one could combine patient care and research if they had the dual MD-PhD. Gerald noted, “The premed adviser at the house [dormitory] was an MD-PhD. He did have a relatively big influence on my decision to pursue MD-PhD.” A dual-degree meant that “I don’t have to give up one side of something that I find exciting and want to explore.”

Participants were motivated by a combination of positive research experiences and personal reasons to pursue the MD-PhD. For example, Debbie shared,

After college I worked as a research technician in a lab studying HIV, and I worked with a lot of physicians who also did research. I sort of liked the idea of the variety in their careers, so I was looking into programs that would allow me to see patients plus do research, and that was how I decided to apply to [the] MD-PhD program.

Personal or family reasons also was a motivation for pursuing MD-PhD. Gerald reasoned, “my grandma was often sick in the nursing home. Going back and forth from the hospital to the nursing home to home. I wanted to help people like her.”

In summary, participants wanted to pursue MD-PhD to be able to work in two worlds—clinical practice and research. Clearly, very positive college research experiences, mentorship, and personal reasons also played big roles in participants’ decisions to pursue the dual MD-PhD degree. And for some, the icing on the cake was the lure of opportunities to participate in a variety of professional activities that they could enjoy as an MD-PhD. So what happened to make these individuals change their minds?

Why participants left the MD-PhD training?

Aaron and Eva left their MD-PhD program at the end of second year without starting the PhD training phase at all; the other five participants completed some of their PhD training before discontinuing the MD-PhD program ( Table 2 ). Once in the program, the influence of positive role models and opportunities that drew candidates to the program was weakened by a variety of factors. Four recurrent themes emerged from the data with regard to participants’ reasons for leaving the MD-PhD program without completing the requirements for both degrees ( Table 3 ), which we describe below.

“Why participants left MD-PhD training?”: Frequency for Each Theme

Declining interest in research

Three participants (Aaron, Debbie, and Eva) shared that although they joined an MD-PhD program to pursue research as well as clinical care, their interest in research and earning a PhD declined shortly after starting the program, which contributed to their decisions to leave the program. At the time of the interview, Aaron was a faculty of clinical research at a medical school and in his sixties (describing experiences from his twenties), and Eva was a second-year medical student in her twenties. Yet, both shared similar experiences of a decline in interest in research following the first few research rotations during their MD-PhD training. Both left their MD-PhD program at the end of their second year of medical school, without formally starting PhD training at all, although both had pursued summer research opportunities during medical school.

For both, it was a combination of being exposed to interesting clinical problems during MD pre-clinical phase, summer rotations shortly after that did not yield research, and a declining interest in research, where “All of a sudden, the PhD just didn’t seem like the thing that I wanted to do anymore, even though when I applied a year and a half ago, I was super excited about it” (Eva). In both cases, lab rotations did not fit research interests, creating doubts about how attractive the PhD would be. Both had an enriching research experience in college that contributed to their decision of doing an MD-PhD. However, once the program started, the excitement:

sort of fizzled. I couldn’t really find something that would keep me interested in that same way. … I was less than thrilled about what I was doing. That was why I first started questioning what I am looking to get out of this. (Eva)

Aaron did not want to put his clinical training on hold after two years and “take off three or four years to go into a lab when I didn’t have a hot project that I was totally enthused about, having had my project from the prior summer sizzle out.” He felt frustrated “not having something [in research] that I had a lot of enthusiasm for. I’d heard about all these fascinating clinical issues and conditions and examined just a couple of patients and thought that was very exciting.” That led him to gravitate towards only the MD degree. The structure of MD-PhD program felt illogical, “giving you the preparation for going into clinics and then saying, ‘Okay, we’ll put that on hold for four years and let’s go do research,’” Aaron shared. At the end of two years, when their MD-PhD cohort split with the rest of the MD classmates, he decided to only continue his clinical training.

This was also largely as a result of positive pre-clinical experiences where both Aaron and Eva learnt a lot from the preceptorship in the first year, an elective mentored experience where one was paired up with a physician to shadow and be involved in doing interviews and physical exams with patients. Debbie, who left after the third year of the MD-PhD program (after two years of medical school and one year of PhD) did that due to the uncertainty of producing research results and lengthy training for the PhD. At the start of her MD-PhD program, she “loved the medical school curriculum and working with other medical students.” However, when she started her PhD training at the beginning of third year, she did not like research as much and felt underprepared for research compared to her MD-PhD peers. She was “leaning more towards medicine” and “didn’t quite fit the MD-PhD profile.” She shared being “not excited everyday by going to the lab, the way I am excited to go to the hospital every day. I just felt like I was missing something. I was unhappy and frustrated doing research.” She realized that she enjoyed clinical training more than research, did not feel as prepared or enthused about getting a PhD by the third year, and felt out of place in the research lab. Like Eva, Debbie would prefer conducting research during residency rather than continuing training for the PhD and ultimately being responsible for running a research lab as a principal investigator.

Isolation and lack of social integration during the different training phases

Social integration broadly describes the ways in which MD-PhD students were able to assimilate into the different cultures during the various training phases. Students described the challenges they experienced and ability to interact with other MD-PhD students as well as with PhD-only and MD-only students during the respective research- and clinical-training phases. Five participants (Ben, Carrie, Eva, Francesca, and Gerald) described challenges in integrating socially in different phases of the MD-PhD program that eventually contributed to their decision of leaving the MD-PhD program. Lack of both family and peer interaction contributed to feelings of isolation.

Family interaction:

There were feelings of isolation due to living far away from family and a cohesive community with which they were familiar. Eva shared that eventually, the novelty of MD-PhD went away and stress related to how long the training was going to take set in. None of the seven participants had an immediate family member in medicine, and four of them were first-generation college students. Having a physician parent might have provided participants with more opportunities and resources to understand and feel comfortable with the demands MD-PhD training. There was a “disconnect in how much my family understands about what I’m doing here at school,” shared Eva. Families sometimes did not understand the academic pressures or the purpose of undergoing such a long training. Although participants reported they did not get much family support while pursuing MD-PhD, Ben shared that he received family support when he decided to leave the program.

Peer interaction:

Isolation due to poor peer interactions started as early as by the second year of MD-PhD training. Socialization opportunities during PhD were inadequate and not as fulfilling, making “the cultural transition from medicine to science a very hard one” (Francesca). It was difficult to mingle with PhD-only students who had already gone through a year of classes and lab rotations with other PhD students and had formed their groups. Participants felt like outsiders in the PhD program. Francesca felt frustrated interacting “with the same five people all day, every day. I was feeling isolated from other people.”

I loved interacting with the patients. I loved the immediacy of medicine. It was a slow realization for me over the last year and a half that I was in the lab that I was much more passionate about the day-to-day work of medicine than I was about the day-to-day work of science. I think part of it as the sort of solitary nature of it [lab research]. I feel like I’m more of a people person than I could be while I was in the lab. (Francesca)

Lack of a social circle was also a challenge, Gerald shared not having “close friends who were doing it [completing MD-PhD training]. Maybe that would have given me more insight into the day-to-day life and might have swayed me a different way.” When a large cohort of MD-PhD students split up to go to different departments during their PhD training, daily interactions with fellow MD-PhD students decreased for him.

Ben felt like being in a difficult environment and a “strange, no-man’s land” to work where neither the MD nor the PhD students considered MD-PhD students one of their own.

[It was like a] cold war between the MD and the PhD faculty at a medical school. The PhDs feel that their degree is of slightly higher rank than an MD and should be treated thus. In a medical school, the MDs insist that [they rank higher]. Both sides feel that they should be in charge, and the other ones are the secondary people. (Ben)

Carrie felt that it would be less stressful if she left PhD training since she had not met a single MD-PhD graduate who was happy. MD students “looked down” on MD-PhD students, considering them to be poor clinicians “because you split your time doing research” and “the PhD did not help in the clinic,” she shared, adding that fellow PhD students did not consider MD-PhDs as serious researchers, saying that MD-PhD students’ “research training was watered down.” Both Carrie and Francesca felt that students in each phase were territorial. Carrie described an “us-against-them mentality”—where MD students considered the PhD-phase of the MD-PhD program as “getting a vacation,” and PhD students were of the opinion that “this isn’t med school where people will hold your hand and spoon-feed you what you need to know.” Francesca felt the cultural transition to the PhD program and the several-year-long gap in medical training were formidable challenges.

In addition to unsatisfactory peer-interactions, Eva eventually realized she enjoyed the daily interactions she experienced working in a hospital more than while conducting research, “which is very much sort of intellectual and introverted. What changed most were the internal factors about what I want out of my career and my life.”

Suboptimal PhD-advising experiences

Three participants (Ben, Carrie, and Gerald) described several challenges related to inadequate mentoring and PhD-advising that contributed to their decision to leave the program. Lack of adequate mentoring during a very regimented MD-PhD training was a widely discussed challenge. Those who left the program described the mentoring they received as minimal, inadequate, sparse, and hands-off. Advisers did not always help in coping with the stress of a long training process, especially during PhD when students had already spent a few years in the program. This was especially discouraging for first-generation students who had received no guidance at home. “Nobody asked if there were problems down there [in the PhD lab] until I did my resignation letter, and then they’re like, ‘Oh, well, what can we do to get you to stay?’ At this point, nothing,” shared Carrie.

Students lacked the bigger picture of what an MD-PhD would be doing ten years down the line, the MD-PhD’s perspective on career development and how to handle training challenges, which could only be provided by MD-PhD advisers (compared to advisers with an MD-only or PhD-only degree). Female MD-PhD students sought female MD-PhD advisers to understand how to achieve work-life balance, who were even rarer to find. Overall, MD-PhD advisers were hard to find.

In addition to bad experiences with PhD advisers and lack of MD-PhD advisers overall, a positive experience with an MD preceptor actually steered students away from a PhD towards an MD-only program. Overall, conflicts arose when adviser and student’s professional goals and values did not match. This happened when a PhD adviser only trained students to become the next generation of principal investigators in a basic science research lab, while that was not the goal for an MD-PhD student. This mismatch made the relationship uncomfortable, especially when advisers “expressed negative opinions of medical students” and treated them more like an employee, shared Gerald.

Ben shared that many PhD advisers were “hostile to the fact that I was an MD-PhD student” and “wore a chip on their shoulder all the time over their position vis-a-vis the doctors. That was just generally a difficult environment to function in.” PhD advisers especially made a difference in a good or bad way because the PhD training process itself was long, with years of research not always yielding publishable results. Given this uncertainty, having young, inexperienced, and pre-tenure PhD advisers further posed challenges, created negative experiences, and discouraged MD-PhD students from completing a PhD. Ben eventually lost his PhD support and was “kicked out against my will for having made inadequate progress” in research. He shared that PhD advisers had “full authority to judge on any criteria they want whether someone has made adequate progress,” and there was no legal defense against that, even if certain committee members did not agree with the decision to expel a student. Often, when a PhD collaboration between faculty and MD-PhD student did not work out, it was difficult to identify another PhD adviser because of smaller MD-PhD programs (compared to PhD-only programs) with fewer available faculty.

When there was lack of MD-PhD advisers, having a better adviser in one phase could disproportionately shift the balance and make students want to complete that part of the training. Gerald had issues working with his PhD adviser, but his MD mentor was very supportive and “willing to meet with me any time to discuss how things are going in medical school, getting back into study habits for medical school after being out for four years.”

Unforeseen obstacles to completing PhD research requirements

Four participants (Ben, Carrie, Francesca, and Gerald) described various unforeseen circumstances that they experienced while completing research requirements during the PhD-phase, which contributed to their decision to leave MD-PhD training. Gerald was in his sixth year (two years of MD and four years of PhD) when he left MD-PhD training. At the beginning of PhD training, none of the lab rotations culminated into a fruitful experience to facilitate completing PhD. Sometimes, “animal models did not work,” forcing one to abandon experiments after many years of effort.

It [the animal model] was still expressing the gene. It was still making the protein, but the phenotypes that…. were no longer there. After several generations of outbreeding were still not there. I was the only person using this model. (Gerald)

Experimental failures created tension between Gerald and committee members because “there was kind of a disconnect between what the rest of my committee expected and what my mentor was able to support.” Even when the program advised to start a new project, it was not possible; Gerald’s PhD adviser “didn’t really have the time or the energy to get that [a new project] off the ground.” Lack of time became a challenge.

I had two weeks to write a completely detailed proposal on this new project. Based on my experience just working with the phenotype and the amount of time and energy that went into that, then looking at [how to] be able to get this new project finished, it would’ve required even more time and energy. It no longer seemed feasible to me. (Gerald)

The possibility of joining a different lab was also eliminated due to time constraints. Francesca, who was in her eighth year of training (two years of MD and six years of PhD) when she left the program, continued to lose more time when the PhD adviser moved to a different university and there were facility-based technical problems.

They constructed [for] us a containment facility instead of a clean room for some of the work, so the airflow was backwards, and all the cultures got contaminated for months. I think I probably lost about nine months with the move and getting all these things straightened out again. (Francesca)

As a result, she felt that the PhD training was tedious and “the things I love about research are sort of hard to vet.” She added, “after six years of doing something, if you’re still spending a lot of time optimizing it, you’re not necessarily learning anything from that. It’s just sort of rote, repetitive work.” Ultimately, the length of time needed to complete MD-PhD training created other personal challenges that contributed to participants’ decisions to discontinue the MD-PhD program. For Ben and Carrie, the lengthy training time, especially during PhD, deterred them from raising a family. Carrie added, “The system penalizes individuals who need to take a break in between their training.”

This study reported results of an analysis of interview data that were collected for a larger qualitative study of training experiences and career decisions made by individuals pursuing biomedical-research careers. Although findings reported here reflect perspectives of only seven individuals who left the dual MD-PhD program before completion, their narratives provide a deeper understanding of reasons for discontinuing training—reasons that have not been captured in surveys ( Ahn et al., 2007 ) or even in large, retrospective, national-cohort studies ( Jeffe et al., 2014a ).

Six of the seven former MD-PhD students finished medical school and completed requirements for the MD degree. Although five left the MD-PhD program within 2-3 years of matriculation, two left after six or more years of training due to extenuating circumstances related to their PhD advisers. Participants’ narratives included details of their reasons for leaving the MD-PhD program. Overall, four recurrent themes emerged from the data, including: 1) declining interest in research, 2) isolation and lack of social integration during the different training phases, 3) suboptimal PhD-advising experiences, and 4) unforeseen obstacles to completing PhD research requirements. Interestingly, analysis of data from 48 then-current MD-PhD students who participated in the larger study also reflected two of the same challenges, specifically, isolation and the lack of social integration during different training phases due to the need to transition between phases, and suboptimal PhD advising; the other challenges experienced by then-current MD-PhD students included a perceived lack of curricular integration as well as cultural differences between the MD and PhD phases of training ( Chakraverty et al., 2018 ). The current study expands upon findings from that earlier study to examine factors that compelled some students to leave their MD-PhD training altogether, and during which phase they left the program.

Despite the small sample size, study findings add to our understanding of the challenges of completing the requirements for the dual MD-PhD as part of a lengthy and disjointed training program. Participants described a complex interplay between students, faculty, and the administrators, resulting in experiencing difficulties with assimilation and immersion into different MD-PhD program cultures through which they transitioned during training. Prior to entering the MD-PhD program, each of these participants reported having had substantial and positive research experiences in college. However, although potentially crucial for decisions to enroll in MD-PhD programs ( Jeffe et al., 2014b ), and even to apply for and be accepted to medical school in general ( Andriole et al., 2015 ), having substantial, positive college research experiences was not enough to keep these participants in the MD-PhD program. Most of those who left PhD training were still interested in pursuing research in the future, but they did not feel the need for a PhD. According to the 2014 NIH Physician-Scientist Workforce Working Group Report ( Feldman, 2014 ), some of the contributors to a leaky workforce include an aging physician-scientist population, long and poorly compensated training, and fewer role models (especially for women and URMs).

For participants who left their MD-PhD training program, the MD-PhD dual degree ultimately did not seem to enhance their career prospects as a researcher; an MD degree alone was deemed sufficient to conduct clinical research. From these participants’ perspectives, what mattered most for research were the grants that people were awarded, the publications, and even faculty appointments. From their point of view, the PhD degree did little to enhance what an MD could offer.

The literature, however, shows benefits of MD-PhD program participation for sustaining and promoting medical students’ intentions to pursue full-time academic-medicine careers ( Jeffe et al., 2008 ) and acquiring full-time faculty appointments ( Andriole & Jeffe, 2016 ). More than half of MD-PhD graduates in a national cohort of medical school matriculants received academic-medicine faculty appointments ( Jeffe et al., 2012 ) compared with only 18% of MD graduates ( Andriole & Jeffe, 2012 ; Jeffe et al., 2012 ). In addition, compared with MD graduates in this same cohort, MD-PhD graduates were more likely to receive each of individual postdoctoral research fellowships (F32) awards, career development (mentored-K) awards, and research project grants (R01) in models controlling for a number of demographic, research related, and academic variables ( Jeffe & Andriole, 2018 ). Moreover, MD-PhD program graduation also has been shown to be a significant mediator of observed racial/ethnic disparities in mentored-K awards in this national cohort ( Andriole et al., 2017 ).

The most prominent finding of the current study is that most participants who left MD-PhD training did so during the PhD-phase. Prior research on PhD-program attrition suggested that PhD training, including in MD-PhD dual-degree programs, was particularly problematic for students who could not integrate well with their peers during this PhD phase of the program ( Golde, 2000 ). Some of the factors related to PhD-program attrition include social isolation ( Ali et al., 2007 ) and the nature of advising, including perceptions of autonomy and relatedness during dissertation ( Burns & Gillespie, 2018 ). Doctoral faculty tend to attribute causes of doctoral-student attrition to student-level factors, often not acknowledging the role of departmental factors ( Gilmore et al., 2016 ). This is despite evidence that the departmental climate and advisers play an important role in their students’ abilities to complete or not complete their training ( Devos et al., 2016 ). Although other factors such as experiencing mental health and adjustment issues due to the impostor phenomenon (where doctoral students feel like intellectual frauds) have not been documented in the literature on doctoral students’ attrition, such factors have influenced student experiences during doctoral training ( Chakraverty, 2019 ; 2020a , 2020b , 2020c ).

Findings from student participants who were still completing requirements for the MD-PhD program identified the importance of more advanced students serving as peer mentors ( Chakraverty et al., 2018 ). Both then-current MD-PhD students in that study and MD-PhD program drop-outs mentioned the critical need for good faculty mentors, and especially MD-PhD mentors who overcame the challenges they faced as students completing MD-PhD training. Both faculty and peer mentors who have faced similar challenges can provide unique insight into what this long and complex training entails ( Chakraverty et al., 2018 ). Notably, none of the participants who dropped out of the MD-PhD program mentioned having supportive peer mentors.

Purposefully building mentoring relationships might help MD-PhD students stay the course during challenging times. Such mentorship groups could involve an MD-PhD student, more advanced MD-PhD students or recent MD-PhD graduates, and faculty, because transitioning between MD-PhD program phases is particularly challenging for these students ( Chakraverty et al., 2018 ).

MD-PhD students who left the program described many challenges assimilating into each program phase due to the disjointed structure of MD-PhD training that did not allow specific program support for socialization and integration. Students transitioning from MD to PhD phases were expected to already know the values and culture of PhD training as well as what was expected of them during PhD training to be able to blend in, something that participants did not always know. Such seamless integration between the different phases was challenging for the MD-PhD students interviewed, but also may require specific integration strategies through re-immersion programs ( Goldberg & Insel, 2013 ) and career-development programs ( Ciampa et al., 2011 ) at each transition. We strongly recommend such academic and socialization strategies to facilitate cultural integration within a program that is as complex as the MD-PhD dual-degree program.

Previous research reported that women and URM students were less likely to be MD-PhD program graduates compared with all other MD program graduates ( Andriole et al., 2008 ); however, in a national cohort study of MD-PhD program matriculants, neither gender nor race/ethnicity were independently associated with overall attrition from MD-PhD training ( Jeffe et al, 2014a ). Our findings show that while evaluating the possible benefits of pursuing the MD-PhD, participants in the present study mostly discussed the disadvantages of a long training time. MD-PhD completion time increased from an average of 6.6 years in 1980 to 8.0 years between 1998 and 2007 ( Brass et al., 2010 ). Such a long training period may itself be a deterrent to program completion, delaying the time to achieve research independence and leading some students to choose clinical practice over research ( Gordon, 2012 ). Notably, however, the time to first R01, the hallmark of research independence, was nearly 2 years shorter from time of graduation for MD-PhD than MD graduates ( Jeffe & Andriole, 2018 ).

Limitations

This was an exploratory study of a very small sample of mostly White individuals who did not complete the MD-PhD program in the U.S. Given the small sample size and homogeneous demography, the findings are not generalizable to the larger MD-PhD student population, in the U.S. or elsewhere. Although age at the time of starting MD-PhD training was not asked, it is evident that most, if not all started their MD-PhD training in their twenties. Further, two participants in their forties and sixties, both medical school faculty when they were interviewed, recalled their experiences in the MD-PhD program from more than a decade before being interviewed, which could be affected by recall bias; however, their experiences were similar to the other participants who only recently left their MD-PhD program when they were interviewed. Nevertheless, findings provide important insight into the reasons for discontinuing their MD-PhD training through a qualitative examination of MD-PhD student narratives, which, to our knowledge, has never before been undertaken. However, the phenomenon of MD-PhD program attrition needs to be examined in greater detail, with a larger and more diverse sample of MD-PhD students who left the program. In addition, we did not elaborate on thematic reasons that were not reported by multiple people, which does not mean that reasons reported by only one person were not important. Nor does it mean that reasons reported only by one person here would not be reported as a recurrent theme had we had interviewed a larger sample of participants who had left the MD-PhD program. Indeed, only one participant, who was Hispanic, described in detail how her decision to leave the program was influenced by the need to stay close to her family and Hispanic community. Although the racial/ethnic diversity of MD-PhD program graduates increased from 5.0% of graduates from URM groups in 1995 to 9.6% in 2015 ( AAMC, 2016 ), URM representation among MD-PhD graduates is considerably lower than their overall representation of more than 30% in the U.S. population ( Colby & Ortman, 2017 ). Additional research is needed to examine URM MD-PhD students’ reasons for MD-PhD program attrition.

Implications and future directions

The findings of this study provide a perspective to understand doctoral research capacity building. While capacity building at the micro-level examines how students transition between the various phases of their training and transform into scholars ( Lovitts, 2005 ), capacity building at the macro-level examines the larger-level trends such as increasing demographic diversity and skill building ( Trostle, 1992 ). Overall, building one’s capacity to be an independent investigator should ideally entail structured mentoring and supervision in the relevant content area, developing specialized, transferable skills, as well as professional development and mentoring to learn about a variety of career prospects outside academia. MD-PhD program attrition can have both micro- and macro-level implications. Micro-level implications include costs to funding agencies and MD-PhD programs ( Jeffe & Andriole, 2011 ; Jeffe et al., 2014a ), as well as to faculty mentors and students themselves (i.e., in time lost and financial burden). It also has macro-level implications in terms of a reduction in the cadre of highly trained, clinical and translational science researchers. Although based on a small sample size, the fact that most attrition happened at the PhD-training level calls for a deeper examination of the challenges students described herein regarding their experiences during the PhD-training phase of MD-PhD training. Findings shed light on situations and experiences that dissuaded these students from completing their PhD training. We urge future research to more deeply examine how interactions among students, faculty and administrators in various settings, such as classrooms, research labs, and clinics, and between different schools and departments, can help MD-PhD students fully integrate into each new program phase they are entering and to continue in the program to completion.

As women and some racial/ethnic groups are underrepresented among MD-PhD program trainees ( Jeffe et al., 2014b ), increasing the diversity of trainees in MD-PhD programs might ultimately serve to increase both the size and diversity of the larger physician-scientist workforce to better meet the needs of an increasingly diverse population ( Milewicz et al., 2015 ; NIH, 2014 ). Examining MD-PhD training experiences through the lens of gender and race/ethnicity should be undertaken in future research with a larger and more diverse sample.

Although greater planned career involvement in research at matriculation was observed to be a predictor of MD-PhD program completion ( Jeffe et al., 2014a ), we found that extenuating circumstances during students’ training in these programs, and apparently, especially during the PhD phase of training, served to derail some of these students’ aspirations to graduate with the MD-PhD dual degree. Attendance at institutions with MSTP funding has been shown to be beneficial and predictive in terms of MD-PhD program completion ( Jeffe et al., 2014a ), and students who attended schools supported by MSTP funds especially benefited during their PhD training ( Goldstein & Brown, 1997 ; Jeffe & Andriole, 2011 ; NIH-NIGMS, 1998 ). However, students whose research is funded solely by their advisers’/mentors’ grants are at greater risk of dropping out of the program for lack of funding, if the advisers’/mentors’ labs closed because they could not renew their grants in the middle of the MD-PhD student’s training in their lab. Institutional MSTP funding has been found to be predictive of students’ retention in the program ( Jeffe & Andriole, 2011 ) and of faculty appointment among MD-PhD graduates ( Andriole & Jeffe, 2016 ).

This paper examined interview responses from seven participants in a larger study who left their MD-PhD programs before completing training; two participants had completed the MD and were academic-medicine faculty, four were completing medical school, and one dropped out of medicine to complete a PhD in Education. Participants reported enrolling in MD-PhD programs to work in both clinical practice and research. Very positive college research experiences, mentorship, and personal reasons played big roles in participants’ decisions to pursue the dual MD-PhD degree. However, once in the program, the influence of earlier positive role models and opportunities that drew candidates to the program was found lacking in the MD-PhD program and weakened their resolve to continue to completion. Four themes emerged as reasons for leaving the MD-PhD program: declining interest in research, isolation and lack of social integration during the different training phases, unsatisfactory PhD-advising/mentoring, and unforeseen obstacles to completing PhD research requirements. We conclude that providing better institutional and social support for the timely completion of research and targeted research mentorship are essential to retaining and promoting the success of students during the PhD phase of their MD-PhD program training. The themes that emerged from participants’ narratives in the current study suggest that targeting interventions to improve students’ educational and research experiences, mentorship, and integration into the different cultures of each program phase are crucial for retention of MD-PhD students through to completion of the program. These same challenges arising from having to transition into different phases of the MD-PhD program were described as well in a larger sample of 68 students who were still in training for the dual MD-PhD degree ( Chakraverty et al., 2018 ). Through a deeper examination of reasons for attrition, MD-PhD programs can find ways to improve training experiences and improve student retention; this can strengthen the biomedical-research-workforce capacity.

Acknowledgment

The authors thank all those who participated in this study and shared their experiences as well as other members of the research team who helped in data collection. This work was supported by a grant from the NIH National Institute of General Medical Sciences (NIGMS) (R01 GM094535). D.B.J. was also supported by NIGMS R01 GM085350. Preliminary findings were presented at the 2017 American Educational Research Association Annual Meeting in San Antonio, TX, and at the 2017 Association of American Medical Colleges Annual Meeting in Boston, MA.

Devasmita Chakraverty , Ph.D., is Assistant Professor at the Ravi J. Matthai Centre for Educational Innovation, Indian Institute of Management Ahmedabad, India. Her research has focused on workforce development in science, technology, engineering, mathematics, and medicine (STEMM). Prior research published in CBE- Life Sciences Education has examined transition experiences of students during MD-PhD training. She also examines reasons why students and professionals experience the impostor phenomenon in STEMM. Dr. Chakraverty has earned a Ph.D. (Science Education) from the University of Virginia, M.P.H. from the School of Public Health, University of Washington, and M.Sc. (Environmental Sciences) from the University of Calcutta (India).

Contributor Information

Devasmita Chakraverty, Indian Institute of Management Ahmedabad, Ahmedabad, India.

Donna B. Jeffe, Washington University in St. Louis, St. Louis, MO, U.S.A.

Katherine P. Dabney, Virginia Commonwealth University, Richmond, VA, U.S.A.

Robert H. Tai, University of Virginia, Charlottesville, VA, U.S.A.

• Ahn J, Watt CD, Man LX, Greeley SAW, & Shea JA (2007). Educating future leaders of medical research: Analysis of student opinions and goals from the MD–PhD SAGE (Students’ Attitudes, Goals, and Education) survey . Academic Medicine , 82 ( 7 ), 633–645. [ PubMed ] [ Google Scholar ]
• Ali A, Kohun F, & Levy Y (2007). Dealing with social isolation to minimize doctoral attrition- A four stage framework . International Journal of Doctoral Studies , 2 ( 1 ), 33–49. [ Google Scholar ]
• Andriole DA, Whelan AJ, & Jeffe DB (2008). Characteristics and career intentions of the emerging MD/PhD workforce . Jama, The Journal of the American Medical Association , 300 ( 10 ), 1165–1173. [ PubMed ] [ Google Scholar ]
• Andriole DA, & Jeffe DB (2016). Predictors of full-time faculty appointment among MD–PhD program graduates: A national cohort study . Medical education online , 21 ( 1 ), 30941. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Andriole DA, & Jeffe DB (2012). The road to an academic medicine career: a national cohort study of male and female US medical graduates . Academic Medicine: Journal of the Association of American Medical Colleges , 87 ( 12 ), 1722–1733. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Andriole DA, Jeffe DB, & Tai RH (2015). Participation in college laboratory research apprenticeships among students considering careers in medicine . Medical Education Online , 20 ( 1 ), 27231. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Andriole DA, Yan Y, & Jeffe DB (2017). Mediators of racial/ethnic disparities in mentored K award receipt among US LCME-accredited medical school graduates . Academic Medicine: Journal of the Association of American Medical Colleges , 92 ( 10 ), 1440–1448. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Antin TM, Constantine NA, & Hunt G (2015). Conflicting discourses in qualitative research: The search for divergent data within cases . Field Methods , 27 ( 3 ), 211–222. [ Google Scholar ]
• Association of American Medical Colleges. AAMC Data Book: Medical Schools and Teaching Hospitals by the Numbers . Washington, DC: Association of American Medical Colleges; 2009. [ Google Scholar ]
• Association of American Medical Colleges. U.S. Medical School Applicants and Students 1982-1983 to 2011-2012 . 2012. Accessed August 4, 2020, from https://www.aamc.org/download/153708/data/charts1982to2012.pdf .
• Association of American Medical Colleges. Table 35. U.S. Medical School M.D.–Ph.D. Graduates by Race/Ethnicity, 1980–2015. In: Diversity in Medical Education: Facts & Figures 2016 . Washington, DC: AAMC. 2016. Accessed August 4, 2020, from http://www.aamcdiversityfactsandfigures2016.org/report-section/section-5/ [ Google Scholar ]
• Bogdan R, & Biklen SK (1997). Qualitative research for education: An Introduction to Theories and Methods , 5th ed.,. Boston, MA: Pearson Allyn & Bacon. [ Google Scholar ]
• Bossé D, Milger K, & Morty RE (2011). Clinician-scientist trainee: A German perspective . Clinical and Investigative Medicine , E324–E329. [ PubMed ] [ Google Scholar ]
• Bourke S, Holbrook A, Lovat T, & Farley P (2004, November). Attrition, completion and completion times of PhD candidates . In AARE annual conference, Melbourne (Vol. 28 ). [ Google Scholar ]
• Brass LF, Akabas MH, Burnley LD, Engman DM, Wiley CA, & Andersen OS (2010). Are MD–PhD programs meeting their goals? An analysis of career choices made by graduates of 24 MD–PhD programs . Academic Medicine: Journal of the Association of American Medical Colleges , 85 ( 4 ), 692–701. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Burns EM, & Gillespie CW (2018). A phenomenological study of attrition from a doctoral cohort program: Changes in feelings of autonomy and relatedness in the dissertation stage . International Journal of Doctoral Studies , 13 , 517–537. [ Google Scholar ]
• The Carnegie Classification of Institutions of Higher Education. (n.d.). Basic classification description . https://carnegieclassifications.iu.edu/classification_descriptions/basic.php
• Chakraverty D (2013). An examination of how women and underrepresented racial/ethnic minorities experience barriers in biomedical research and medical programs [Doctoral dissertation. University of Virginia; ]. https://digitalcommons.unl.edu/dberspeakers/43 [ Google Scholar ]
• Chakraverty D (2019). Impostor phenomenon in STEM: Occurrence, attribution, and identity . Studies in Graduate and Postdoctoral Education , 10 ( 1 ), 2–20. 10.1108/SGPE-D-18-00014 [ CrossRef ] [ Google Scholar ]
• Chakraverty D (2020a). PhD student experiences with the impostor phenomenon in STEM . International Journal of Doctoral Studies , 15 ( 1 ), 159–180. 10.28945/4513 [ CrossRef ] [ Google Scholar ]
• Chakraverty D (2020b). The impostor phenomenon among postdoctoral trainees in STEM: A US-based mixed-methods study . International Journal of Doctoral Studies , 15 , 329–352. 10.28945/4589 [ CrossRef ] [ Google Scholar ]
• Chakraverty D (in press; 2020c). The impostor phenomenon among Black doctoral and postdoctoral scholars in STEM . International Journal of Doctoral Studies [ Google Scholar ]
• Chakraverty D, Jeffe DB, & Tai RH (2018). Transition experiences in MD–PHD programs . CBE—Life Sciences Education , 17 ( 3 ), ar41. 10.1187/cbe.17-08-0187 [ PMC free article ] [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Cohen D, & Crabtree B (2006). Qualitative research guidelines project . Accessed August 4, 2020, from http://www.qualres.org/HomeSemi-3629.html
• Colby SL, & Ortman JM (2017). Projections of the size and composition of the US population: 2014 to 2060: Population estimates and projections . Accessed August 4, 2020, from http://wedocs.unep.org/bitstream/handle/20.500.11822/20152/colby_population.pdf?sequence=1
• Ciampa EJ, Hunt AA, Arneson KO, Mordes DA, Oldham WM, Vin Woo K, … & Dermody TS (2011). A workshop on leadership for MD/PhD students . Medical Education Online , 16 ( 1 ), 7075. Accessed August 4, 2020, from https://www.tandfonline.com/doi/full/10.3402/meo.v16i0.7075 [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Dabney KP, Chakraverty D, Hutton AC, Warner KA, & Tai RH (2016). The bachelor’s to PhD transition: Factors influencing PhD completion among women in chemistry and physics . Bulletin of Science, Technology & Society , 36 ( 4 ), 203–210. 10.1177/0270467617710852 [ CrossRef ] [ Google Scholar ]
• Devos C, Boudrenghien G, Van der Linden N, Frenay M, Azzi A, Galand B, & Klein O (2016). Misfits between doctoral students and their supervisors:(how) are they regulated? . International Journal of Doctoral Studies , 11 , 467–486. [ Google Scholar ]
• DiCicco-Bloom B, & Crabtree BF (2006). The qualitative research interview . Medical education , 40 ( 4 ), 314–321. [ PubMed ] [ Google Scholar ]
• dos Santos Rocha A, Scherlinger M, Ostermann L, Mehler DMA, Nadiradze A, Schulze F, … & Kijlstra JD (2020). Characteristics and opinions of MD-PhD students and graduates from different European countries: a study from the European MD-PhD Association . Swiss Medical Weekly , 150 ( 1516 ). [ PubMed ] [ Google Scholar ]
• Faulkner SL, & Trotter SP (2017). Data saturation . The international encyclopedia of communication research methods , 1–2. [ Google Scholar ]
• Feldman AM (2014). The National Institutes of Health Physician-Scientist Workforce Working Group report: a roadmap for preserving the physician-scientist . Clinical and translational science , 7 ( 4 ), 289. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Garrison G, Mikesell C, & Matthew D (2007). Medical school graduation and attrition rates . AAMC Analysis in Brief , 7 ( 2 ), 1–2. Accessed August 4, 2020, from https://www.aamc.org/download/102346/data/aibvol7no2.pdf [ Google Scholar ]
• Geiger RL (1997). Doctoral education: The short-term crisis vs. long-term challenge . The Review of Higher Education , 20 ( 3 ), 239–251. [ Google Scholar ]
• Gilmore J, Wofford AM, & Maher MA (2016). The flip side of the attrition coin: Faculty perceptions of factors supporting graduate student success . International Journal of Doctoral Studies , 11 , 419–439. [ Google Scholar ]
• Glaser BG, & Strauss AL (1967). The discovery of grounded theory: strategies for qualitative research . New York: Aldine de Gruyter. [ Google Scholar ]
• Goldberg C, & Insel PA (2013). Preparing MD-PhD students for clinical rotations: navigating the interface between PhD and MD training . Academic Medicine: Journal of the Association of American Medical Colleges , 88 ( 6 ), 745–747. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Golde CM (2000). Should I stay or should I go? Student descriptions of the doctoral attrition process . The Review of Higher Education , 23 ( 2 ), 199–227. [ Google Scholar ]
• Goldstein JL, & Brown MS (1997). The clinical investigator: bewitched, bothered, and bewildered--but still beloved . The Journal of Clinical Investigation , 99 ( 12 ), 2803–2812. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Goldstein MJ, & Kohrt HE (2012). What happened to the concept of the physician–scientist? . Academic Medicine , 87 ( 2 ), 132–133. [ PubMed ] [ Google Scholar ]
• Gordon R (2012). The vanishing physician scientist: a critical review and analysis . Accountability in Research , 19 ( 2 ), 89–113. [ PubMed ] [ Google Scholar ]
• Jeffe DB, & Andriole DA (2018). Prevalence and predictors of US medical graduates’ federal F32, mentored-K, and R01 awards: a national cohort study . Journal of Investigative Medicine , 66 ( 2 ), 340–350. doi: 10.1136/jim-2017-000515. [ PMC free article ] [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Jeffe DB, & Andriole DA (2011). A national cohort study of MD–PhD graduates of medical schools with and without funding from the National Institute of General Medical Sciences’ Medical Scientist Training Program . Academic Medicine: Journal of the Association of American Medical Colleges , 86 ( 8 ), 953–961. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Jeffe DB, Andriole DA, Hageman HL, & Whelan AJ (2008). Reaping what we sow: the emerging academic medicine workforce . Journal of the National Medical Association , 100 ( 9 ), 1026–1034 [ PubMed ] [ Google Scholar ]
• Jeffe DB, Andriole DA, Wathington HD, & Tai RH (2014a). Educational outcomes for MD-PhD program matriculants: a national cohort study . Academic Medicine: Journal of the Association of American Medical Colleges , 89 ( 1 ), 84–93. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Jeffe DB, Andriole DA, Wathington HD, & Tai RH (2014b). The emerging physician-scientist workforce: demographic, experiential, and attitudinal predictors of MD-PhD program enrollment . Academic Medicine: Journal of the Association of American Medical Colleges , 89 ( 10 ), 1398–1407. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Jeffe DB, Yan Y, & Andriole DA (2012). Do research activities during college, medical school, and residency mediate racial/ethnic disparities in full-time faculty appointments at US Medical schools? Academic Medicine: Journal of the Association of American Medical Colleges , 87 ( 11 ), 1582–1593. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Jones AA, Ng E, Deguise MO, Mak L, Ouyang B, Sivapragasam M, … & Wang X (2016). MD/PhD Training in Canada: Results from a national trainee and program director review . Clinical and Investigative Medicine , E132–E139. Available at: https://cimonline.ca/index.php/cim/article/view/27092 . Accessed August 4, 2020. [ PubMed ] [ Google Scholar ]
• Kersbergen CJ, Bowen CJ, Dykema AG, Koretzky MO, Tang O, & Beach MC (2020). Student perceptions of MD-Ph. D. programs: A qualitative identification of barriers facing prospective MD-PhD applicants . Teaching and Learning in Medicine , 32 ( 1 ), 1–10. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Kong X, Chakraverty D, Jeffe DB, Andriole DA, Wathington HD, & Tai RH (2013). How do interaction experiences influence doctoral students’ academic pursuits in biomedical research? Bulletin of Science, Technology & Society , 33 ( 3-4 ), 76–84. 10.1177/0270467613516754 [ PMC free article ] [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Kuehnle K, Winkler DT, & Meier-Abt PJ (2009). Swiss national MD-PhD program: an outcome analysis . Swiss medical weekly , 139 ( 37 ), 540–546. [ PubMed ] [ Google Scholar ]
• Kvale S, & Brinkmann S (2009). Interviews: Learning the craft of qualitative research interviewing (2nd Ed.). Thousand Oaks, CA: Sage Publications, Inc. [ Google Scholar ]
• Kwan JM, Daye D, Schmidt ML, Conlon CM, Kim H, Gaonkar B, … & Winter KQ (2017). Exploring intentions of physician-scientist trainees: factors influencing MD and MD/PhD interest in research careers . BMC Medical Education , 17 ( 1 ), 115. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Lott JL, Gardner S, & Powers DA (2009). Doctoral student attrition in the STEM fields: An exploratory event history analysis . Journal of College Student Retention: Research, Theory & Practice , 11 ( 2 ), 247–266. [ Google Scholar ]
• Lovitts BE (2005). Being a good coursefvtaker is not enough: a theoretical perspective on the transition to independent research . Studies in Higher Education , 30 ( 2 ), 137–154. [ Google Scholar ]
• Marshall C, & Rossman GB (2006). Designing qualitative research (4th ed.). Sage Publications. [ Google Scholar ]
• Miles MB, & Huberman AM (1994). Qualitative data analysis: An expanded sourcebook . Sage Publications. [ Google Scholar ]
• Michalski DS, Cope C, & Fowler GA (2016). Summary Report: Student Attrition. Graduate Study in Psychology 2016 . American Psychological Association . [ Google Scholar ]
• Milewicz DM, Lorenz RG, Dermody TS, & Brass LF (2015). Rescuing the physician-scientist workforce: The time for action is now . The Journal of Clinical Investigation , 125 ( 10 ), 3742–3747. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• National Institute of General Medical Sciences (US). (1998). The careers and professional activities of graduates of the NIGMS medical scientist training program (No. 98) . US Department of Health and Human Services, Public Health Service, National Institutes of Health, National Institute of General Medical Sciences. Accessed August 4, 2020, from, http://publications.nigms.nih.gov/reports/mstpstudy/#10 [ Google Scholar ]
• National Institute of General Medical Sciences. Medical Scientist Training Program (MSTP) institutions . For award period 7/1/2019–6/30/2020 . At: http://www.nigms.nih.gov/Training/InstPredoc/PredocInst-MSTP.htm . Accessed August 4, 2020. [ Google Scholar ]
• National Institutes of Health (June, 2014). Physician–Scientist Workforce Working Group report . Accessed August 4, 2020, from http://acd.od.nih.gov/reports/PSW_Report_ACD_06042014.pdf . [ PMC free article ] [ PubMed ]
• Okahana H, Allum J, Felder PP, & Tull G (2016). Implications for practice and research from Doctoral Initiative on Minority Attrition and Completion . Washington, DC: Council of Graduate Schools. [ Google Scholar ]
• Pope C, Ziebland S, & Mays N (2000). Qualitative research in health care: Analysing qualitative data . BMJ: British Medical Journal , 320 ( 7227 ), 114. 10.1136/bmj.320.7227.114 [ PMC free article ] [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Sadler GR, Lee HC, Lim RSH, & Fullerton J (2010). Recruitment of hard-to-reach population subgroups via adaptations of the snowball sampling strategy . Nursing & Health Sciences , 12 ( 3 ), 369–374. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Tai RH, Kong X, Mitchell CE, Dabney KP, Read DM, Jeffe DB, … & Wathington HD (2017). Examining summer laboratory research apprenticeships for high school students as a factor in entry to MD/PhD programs at matriculation . CBE—Life Sciences Education , 16 ( 2 ), ar37. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Thomas DR (2006). A general inductive approach for analyzing qualitative evaluation data . American Journal of Evaluation , 27 ( 2 ), 237–246. [ Google Scholar ]
• Tinto V (1987). Leaving college: Rethinking the causes and cures of student attrition . University of Chicago Press, 5801 S. Ellis Avenue, Chicago, IL: 60637. [ Google Scholar ]
• Trostle J (1992). Research capacity building in international health: definitions, evaluations and strategies for success . Social Science & Medicine , 35 ( 11 ), 1321–1324. [ PubMed ] [ Google Scholar ]
• Twa DD, Skinnider MA, Squair JW, & Lukac CD, on behalf of the Canadian MD/PhD Program Investigation Group. (2017). Cross-sectional-derived determinants of satisfaction with physician-scientist training among Canadian MD/PhD graduates . PloS one , 12 ( 9 ), e0185218. 10.1371/journal.pone.0185218 [ PMC free article ] [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Varki A, & Rosenberg LE (2002). Emerging opportunities and career paths for the young physician-scientist . Nature Medicine , 8 ( 5 ), 437–439. [ PubMed ] [ Google Scholar ]

Doctoral Completion & Time-to-degree

This page contains information about degree conferrals, time-to-degree, and retention for doctoral research programs at Stanford. While the most common academic doctoral degree across the university is the PhD, the JSD in Law and the DMA in Musical Arts are also included here. The MD and JD are considered to be professional degrees and are not included. In all cases below, the academic years reported are "summer start years", meaning that the academic year encompasses a period from the start of the summer term through the end of the following spring term. Please see the  definitions  below the dashboard for more details about how the various metrics presented here are calculated.

More information is available about  doctoral program enrollment and demographics , as well as  doctoral program admissions . Note that local variation in policy and practice regarding admission, matriculation, and degree conferral may affect the departmental and school-level metrics below.

Methodology & Definitions

Time-to-degree.

Time-to-degree is the length of time in years from the first day of the student's first term of enrollment in their doctoral program to the day of their degree conferral. Time-to-degree measures elapsed time only, not enrolled time. It does not stop and start if a student takes a leave of absence. If a student was enrolled in a master's degree program prior to matriculating in the doctoral program the separate time in the master's program is not included even if it was in the same department as the doctoral program. For this reason, time-to-degree may be lower in some doctoral programs where it is common to require completion of a master's degree prior to matriculation in the doctoral program. If a student switches between doctoral programs, time-to-degree is restarted from the first term of enrollment in the new program. The only exceptions to this restart of the clock are when program changes are the result of departmental name changes or other restructuring, or when the new program has the same CIP code as the original program.

Graduation Rate

As with time-to-degree, the start of the 6-year period used to calculate graduation rates is the first term in which the student is enrolled in a doctoral program, regardless of any prior or concurrent enrollment in a master's program. The 6-year rate is based on elapsed time only, not enrolled time. It is not based on the concept of a cohort year or graduation year, but on the actual matriculation term and degree conferral term. For example, if a student enrolled at the start of Spring 2010 and graduated at the end of Winter 2016, they would count towards the 6-year rate; however, if they instead graduated at the end of Spring 2016, their time to degree would be more than 6 years due to the extra term of enrollment.

Degree Conferrals

Numbers of degree conferrals are reported by summer start year. For example, all degree conferred from Summer 2016 through the following Spring 2017 would be reported under the 2016-17 year.

Entering Cohort Status

An entering cohort consists of all students entering a doctoral program during autumn, winter, or spring quarter of a single academic year, as well as those entering during the preceding summer. Students are considered to be current in their program if they are still actively pursuing that degree or are on an approved temporary leave of absence. "Current students in a different PhD program" are students who were enrolled at one point in the selected program but subsequently moved to another doctoral program at Stanford and are still engaged in doctoral study. Students who are listed as "completed" have successfully conferred their degree in the selected program or, if they have completed a different doctoral program, have changed programs and been awarded a doctoral degree by another program at Stanford. Program changes resulting from department name changes, organizational restructuring, or between programs with the same CIP code are not considered "changes" in this context. Students who are shown as "discontinued" have either left the university without a degree or switched to a non-doctoral degree program (in many cases a master's degree).

Visit the  Graduate Admissions website  for more information about pursuing graduate study at Stanford.

The data are available for download in Google Drive .

• Data Source(s): PeopleSoft Campus Solutions, Institutional Research & Decision Support

Stanford University is committed to providing an online environment that is accessible to everyone, including individuals with disabilities. If you cannot access this content or use any features on this site, please contact  [email protected]  to obtain alternate formats.

You may submit feedback on this dashboard through the  feedback form .

PhD completion: an evidence-based guide for students, supervisors and universities

Senior Lecturer in Management, Fellow of the APS College of Organisational Psychologists, Swinburne University of Technology

Disclosure statement

Timothy Colin Bednall does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.

Swinburne University of Technology provides funding as a member of The Conversation AU.

View all partners

Many students enrol in a Master or PhD postgraduate research degree, but few complete them. From 2010-2016 , 437,030 domestic and international students enrolled in postgraduate research programs in Australian public universities. Only 65,101 completed within the same six year period.

This discrepancy does not necessarily mean postgraduate research students “failed” their degree. Common reasons not completing a degree include changes of career goals, work-family conflicts, poor health or financial strain. Alternatively, some students remain enrolled in their degree for long periods without making significant progress.

Even so, the discrepancy is large enough for universities to be concerned. Nobody wants a student to suffer through years of hard work and frustration without achieving their goal.

What does research say about completion rates?

Research has identified several factors that make students more likely to persist with their degrees. These factors are related to the students themselves, their supervisor, and the university environment.

Psychological studies of postgraduate students find the more successful ones tend to perceive themselves as competent and be intrinsically motivated . These are students who enjoy their topic area, perceive their postgraduate studies as a valuable learning experience, and who strongly identify with being a career researcher. Students who are motivated by external factors (such as pursuing a prestigious academic role) are more likely to say they want to quit.

Scholarship holders are more likely to complete their degrees. This is likely because they are academically stronger than non-scholarship holders and are less vulnerable to financial strain. Students can support themselves financially through teaching, research assistant roles or other work, but this must be balanced carefully. Part-time students are less likely to complete their degrees.

Students’ field of study also affects completion rates. A higher proportion of students in sciences tend to complete their degrees than those in arts and humanities. This is likely because students working in the sciences are more often involved in laboratory-based work in teams, where there is greater social support and knowledge exchange. People studying humanities more often work on their research alone.

A positive student-supervisor working relationship is critical. A good supervisor should be an expert in the student’s subject of choice and a supportive mentor. They should help the student navigate through the frustrations and uncertainties of writing a thesis, and help students adjust to the world of academia.

Students are also more likely to finish their research degrees if they have strong connections with their peers . Such connections help students develop their professional identity as researchers, as well as providing opportunities for social support and informal learning .

The quality of associated coursework is also important. Ideally, postgraduate programs should provide students with a sound foundation of research skills and content knowledge, and facilitate ongoing communication with their faculty.

Involvement in formal and informal professional activities is also important. Students who complete tend to participate in departmental events, such as research seminars and professional development workshops. They also tend to participate in academic conferences. These events allow students to learn and expand their networks.

What students and their supervisors should do

First, given the importance of the student-supervisor relationship, universities can provide advice to students about locating and approaching a suitable supervisor. Specifically, students should consider the research area they wish to work in and locate a supervisor with relevant expertise. They should approach supervisors with an openness to negotiating a research topic.

Read more: Ten types of PhD supervisor relationships – which is yours?

Both students and supervisors should be upfront about their expectations about how the supervision will work. An excellent starting point for discussion is the Expectations in Supervision questionnaire. Students and supervisors sometimes have mismatched expectations about how often they should meet, the amount of feedback the supervisor should provide on drafts, and how much counselling and emotional support the supervisor should provide.

Supervisors have an important role in providing a realistic preview of academic life. One useful exercise is to review an academic competency model, such as the Vitae Researcher Development Framework , to discuss which skills academics need. In addition to knowledge of their topic area and research methods, academics increasingly need to be good at managing complex projects, working in multidisciplinary teams, and engaging with industry and media.

This discussion should enable supervisors and students to plan how students will develop their capabilities. Alternatively, it could prompt some students to opt out of a research degree if they think an academic role is not compatible with their goals.

What universities should do

As well as providing research training, universities can also increase the capabilities of students by helping them understand self-handicapping patterns. These include busyness, procrastination and disorganisation.

Students can be guided to replace these with more helpful actions such as scheduling dedicated writing time, reframing difficult tasks as learning opportunities, and developing a work routine. This could be done as part of a workshop or supervisory relationship.

Universities should also encourage greater connectedness between research students to build social support. This could be accomplished through team-based activities or face-to-face events.

For instance, some universities offer Three Minute Thesis , a research communication competition where students present their work in under 180 seconds.

Some universities organise Shut Up and Write sessions, which turns writing into a social experience and limits distractions. These activities can be complemented by encouraging students to become involved in supportive online communities and blogging .

Read more: The rise of writing events gives PhD students the support often lacking in universities

Finally, universities should be dedicated to helping academics develop as supervisors through ongoing training and coaching. Departments could consider tracking the progression of students and ensuring supervisors have the time and skills to take on new students.

Completing a dissertation can be richly rewarding, but it’s the endpoint of a process that’s often long, frustrating and uncertain. Helping students achieve their research aspirations makes academic life a better experience for all involved.

• Postgraduate degrees
• PhD supervisors
• PhD students

Faculty of Law - Academic Appointment Opportunities

Operations Manager

Senior Education Technologist

Audience Development Coordinator (fixed-term maternity cover)

Lecturer (Hindi-Urdu)

Advertisement

Analyzing the relation among different factors leading to Ph.D. dropout using numerical association rule mining

• Published: 07 November 2023
• Volume 29 , pages 375–399, ( 2024 )

Cite this article

• Manevpreet Kaur 1 ,
• Madanjit Singh   ORCID: orcid.org/0000-0002-1151-6729 1 &
• Munish Saini 2  

232 Accesses

Explore all metrics

This article has been updated

Ph.D. dropout is a persistent and challenging issue in higher education, with significant implications for individual students, academic institutions, and the broader society. This research paper aims to explore the factors contributing to Ph.D. students' decision to dropout and their interrelationships. For this purpose, we employed the hybrid topic modeling Bidirectional Encoder Representations from Transformers – Latent Dirichlet Allocation (BERT-LDA) algorithm and Numerical Association Rule Mining (NARM) using a genetic algorithm in QuantMiner. We identified and analyzed individual, institutional, and social factors that affect Ph.D. students in leaving their current degrees. The results suggest that financial constraints, inadequate academic preparation, poor mentoring, social isolation, lack of social support, family responsibilities, and work-life balance are significant elements responsible for dropout. These findings also reveal that these factors are interrelated, and their effects can be mitigated by the academic institution's policies and culture. The outcomes of the study have implications for academic institutions, policymakers, and researchers, who can use them to develop evidence-based strategies and interventions that enhance Ph.D. students' retention and success.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price includes VAT (Russian Federation)

Instant access to the full article PDF.

Rent this article via DeepDyve

Institutional subscriptions

Similar content being viewed by others

Student Dropout Model Based on Logistic Regression

From Transcripts to Insights for Recommending the Curriculum to University Students

Thong Le Mai, Minh Thanh Chung, … Nam Thoai

Investigating school absenteeism and refusal among Australian children and adolescents using Apriori association rule mining

Umme Marzia Haque, Enamul Kabir & Rasheda Khanam

Data availability

All the datasets generated and analyzed during the current study are available from the corresponding author upon reasonable request.

Code availability

Not applicable.

Change history

06 december 2023.

Springer Nature’s version of this paper was updated to rectify the incorrect version of the article PDF proof.

https://figshare.com/s/74a5ea79d76ad66a8af8

https://pypi.org/project/googletrans/

Ali, A., & Gregg Kohun, F. (2006). Dealing with isolation feelings in is doctoral programs. International Journal of Doctoral Studies, 1, 021–033. 10.28945/58

Ampaw, F. D., & Jaeger, A. J. (2011). Understanding the factors affecting degree completion of doctoral women in the Science and Engineering Fields. New Directions for Institutional Research, 2011(152), 59–73. https://doi.org/10.1002/ir.409

Aulck, L., Velagapudi, N., Blumenstock, J., & West, J. (2016). Predicting student dropout in higher education. arXiv .  https://doi.org/10.48550/arXiv.1606.06364

Austin, A. E. (2002). Preparing the next generation of faculty: Graduate school as socialization to the academic career. The journal of higher education, 73 (1), 94–122. https://doi.org/10.1080/00221546.2002.11777132

Article   Google Scholar  

Bair, C. R., & Haworth, J. G. (2005). Doctoral student attrition and persistence: A meta-synthesis of research. Higher education: Handbook of theory and research , 481-534. https://doi.org/10.1007/1-4020-2456-8

Baker, V. L., & Lattuca, L. R. (2010). Developmental networks and learning: Toward an interdisciplinary perspective on identity development during doctoral study. Studies in higher education, 35 (7), 807–827. https://doi.org/10.1080/03075070903501887

Bean, J. P., & Eaton, S. B. (2000). A psychological model of college student retention. Reworking the student departure puzzle, 1 , 48–61.

Google Scholar  

Blei, D. M., Ng, A. Y., & Jordan, M. I. (2003). Latent dirichlet allocation. Journal of machine Learning research, 3 (Jan), 993–1022.

Blyth, S. (1994). Karl Pearson and the correlation curve. International Statistical Review/Revue Internationale de Statistique , 393-403. https://doi.org/10.2307/1403769

Castellanos, J., & Gloria, A. M. (2007). Research considerations and theoretical application for best practices in higher education: Latina/os achieving success. Journal of Hispanic Higher Education, 6 (4), 378–396. https://doi.org/10.1177/1538192707305347

Chai, X., Tang, G., Wang, S., Peng, R., Chen, W., & Li, J. (2020). Deep learning for regularly missing data reconstruction. IEEE Transactions on Geoscience and Remote Sensing, 58 (6), 4406–4423. https://doi.org/10.1109/TGRS.2020.2963928

Chen, Y., Dong, G., Han, J., Wah, B. W., & Wang, J. (2002). Multi-dimensional regression analysis of time-series data streams. In VLDB'02: Proceedings of the 28th International Conference on Very Large Databases (pp. 323–334). Morgan Kaufmann. https://doi.org/10.1016/B978-155860869-6/50036-6

Cyranoski, D., Gilbert, N., Ledford, H., Nayar, A., & Yahia, M. (2011). Education: The phd factory. Nature, 472 (7343), 276–279. https://doi.org/10.1038/472276a276-280 .

Devlin, J., Chang, M. W., Lee, K., & Toutanova, K. (2018). Bert: Pre-training of deep bidirectional transformers for language understanding. arXiv preprint arXiv:1810.04805 . https://doi.org/10.48550/arXiv.1810.04805

Fuhrmann, C. N., Halme, D. G., O’sullivan, P. S., & Lindstaedt, B. (2011). Improving graduate education to support a branching career pipeline: Recommendations based on a survey of doctoral students in the basic biomedical sciences. CBE—Life Sciences Education, 10 (3), 239–249. https://doi.org/10.1187/cbe.11-02-0013

Gardner, S. K. (2009a). Conceptualizing success in doctoral education: Perspectives of faculty in seven disciplines. The Review of Higher Education, 32 (3), 383–406. https://doi.org/10.1353/rhe.0.0075

Gardner, S. K. (2009b). Student and faculty attributions of attrition in high and low-completing doctoral programs in the United States. Higher education, 58 , 97–112. https://doi.org/10.1007/s10734-008-9184-7

George, L., & Sumathy, P. (2023). An integrated clustering and BERT framework for improved topic modeling. International Journal of Information Technology , 1–9. https://doi.org/10.1007/s41870-023-01268-w

Golde, C. M. (2000). Should I Stay or Should I Go? Student Descriptions of the Doctoral Attrition Process. The Review of Higher Education, 23 (2), 199–227. https://doi.org/10.1353/rhe.2000.0004

Golde, C. M. (2005). The Role of the Department and Discipline in Doctoral Student Attrition: Lessons from Four Departments. The Journal of Higher Education, 76 (6), 669–700. https://doi.org/10.1080/00221546.2005.11772304

Griffiths, T. L., & Steyvers, M. (2004). Finding scientific topics. Proceedings of the National academy of Sciences, 101 (suppl_1), 5228–5235. https://doi.org/10.1073/pnas.030775210

Hauke, J., & Kossowski, T. (2011). Comparison of values of Pearson's and Spearman's correlation coefficients on the same sets of data. Quaestiones geographicae, 30 (2), 87–93. https://doi.org/10.2478/v10117-011-0021-1

Heckman, J. J., & LaFontaine, P. A. (2010). The American high school graduation rate: Trends and levels. The review of economics and statistics, 92 (2), 244–262. https://doi.org/10.1162/rest.2010.12366

Huang, G. (2021). Missing data filling method based on linear interpolation and lightgbm. In Journal of Physics: Conference Series (Vol. 1754, No. 1, p. 012187). IOP Publishing. https://doi.org/10.1088/1742-6596/1754/1/012187

Hunter, K. H., & Devine, K. (2016). Doctoral students’ emotional exhaustion and intentions to leave academia. International Journal of doctoral studies, 11 (2), 35–61. https://doi.org/10.28945/3396

Jaromczyk, J. W., & Toussaint, G. T. (1992). Relative neighborhood graphs and their relatives. Proceedings of the IEEE, 80 (9), 1502–1517. https://doi.org/10.1109/5.163414

Kaur, M., & Saini, M. (2023). Indian government initiatives on cyberbullying: A case study on cyberbullying in Indian higher education institutions. Education and Information Technologies, 28 (1), 581–615. https://doi.org/10.1007/s10639-022-11168-4

Kotsiantis, S. (2009). Educational data mining: a case study for predicting dropout-prone students. International Journal of Knowledge Engineering and Soft Data Paradigms, 1 (2), 101–111. https://doi.org/10.1504/IJKESDP.2009.022718

Koufakou, A. (2023). Deep learning for opinion mining and topic classification of course reviews. Education and Information Technologies , 1-25. https://doi.org/10.1007/s10639-023-11736-2

Latif, A., Choudhary, A. I., & Hammayun, A. A. (2015). Economic effects of student dropouts: A comparative study. Journal of global economics . https://doi.org/10.4172/2375-4389.1000137

Levecque, K., Anseel, F., De Beuckelaer, A., Van der Heyden, J., & Gisle, L. (2017). Work organization and mental health problems in PhD students. Research policy, 46 (4), 868–879. https://doi.org/10.1016/j.respol.2017.02.008

Litalien, D., & Guay, F. (2015). Dropout intentions in PhD studies: A comprehensive model based on interpersonal relationships and motivational resources. Contemporary Educational Psychology, 41 , 218–231. https://doi.org/10.1016/j.cedpsych.2015.03.004

Liu, Y., Ott, M., Goyal, N., Du, J., Joshi, M., Chen, D., ... & Stoyanov, V. (2019). Roberta: A robustly optimized bert pretraining approach. arXiv preprint arXiv:1907.11692 . https://doi.org/10.48550/arXiv.1907.11692

Lovitts, B. E. (2002). Leaving the ivory tower: The causes and consequences of departure from doctoral study . Rowman & Littlefield Publishers.

Lovitts, B. E., & Nelson, C. (2000). The hidden crisis in graduate education: Attrition from Ph. D. programs. Academe, 86 (6), 44. https://doi.org/10.2307/40251951

Li, J., Koedel, C., & Zhang, L. (2021). Mitigating Ph.D. dropout rates: Evidence from a randomized control trial. Economics of Education Review, 80 , 102129. https://doi.org/10.1016/j.econedurev.2020.102129

Luengo, J., García-Gil, D., Ramírez-Gallego, S., García, S., & Herrera, F. (2020). Big data preprocessing . Springer. https://doi.org/10.1007/978-3-030-39105-8

Book   Google Scholar  

Lunardi, A. (2009). Interpolation theory (Vol. 9). Pisa: Edizioni della normale.

Maharana, K., Mondal, S., & Nemade, B. (2022). A review: Data pre-processing and data augmentation techniques. Global Transitions Proceedings. https://doi.org/10.1016/j.gltp.2022.04.020

Mason, M. A., & Goulden, M. (2004). Do babies matter (part II)? closing the baby gap. Academe, 90(6), 10. https://doi.org/10.2307/40252699

Mata, J., Alvarez, J. L., & Riquelme, J. C. (2002). Discovering numeric association rules via evolutionary algorithm. In Pacific-Asia conference on knowledge discovery and data mining (pp. 40-51). Springer, Berlin, Heidelberg. https://doi.org/10.1007/3-540-47887-6_5

Moslehi, F., & Haeri, A. (2020). A genetic algorithm-based framework for mining quantitative association rules without specifying minimum support and minimum confidence. Scientia Iranica, 27 (3), 1316–1332.

Nagao, M., & Seki, H. (2016). On mining quantitative association rules from multi-relational data with FCA. In 2016 IEEE 9th International Workshop on Computational Intelligence and Applications (IWCIA) (pp. 81–86). IEEE.  https://doi.org/10.1109/IWCIA.2016.7805753

Noor, M. N., Yahaya, A. S., Ramli, N. A., & Al Bakri, A. M. M. (2014). Filling missing data using interpolation methods: Study on the effect of fitting distribution (Vol. 594, pp. 889–895). Trans Tech Publications Ltd. https://doi.org/10.4028/www.scientific.net/KEM.594-595.889

Nerad, M., & Cerny, J. (1999). Postdoctoral patterns, career advancement, and problems. Science, 285 (5433), 1533–1535. https://doi.org/10.1126/science.285.5433.1533

Nettleton, D. F., Orriols-Puig, A., & Fornells, A. (2010). A study of the effect of different types of noise on the precision of supervised learning techniques. Artificial intelligence review, 33 , 275–306. https://doi.org/10.1007/s10462-010-9156-z

Pal, S. (2012). Mining educational data using classification to decrease dropout rate of students. arXiv preprint arXiv :1206.3078 . https://doi.org/10.48550/arXiv.1206.3078

Paul, P. C., Uddin, M. S., Ahmed, M. T., Hoque, M. M., & Rahman, M. (2022). Semantic Topic Extraction from Bangla News Corpus Using LDA and BERT-LDA. In 2022 25th International Conference on Computer and Information Technology (ICCIT) (pp. 512-516). IEEE. https://doi.org/10.1109/ICCIT57492.2022.10055173

Pérez, B., Castellanos, C., & Correal, D. (2018). Predicting student drop-out rates using data mining techniques: A case study. In Applications of Computational Intelligence: First IEEE Colombian Conference, ColCACI 2018, Medellín, Colombia, May 16-18, 2018, Revised Selected Papers 1 (pp. 111-125). Springer International Publishing. https://doi.org/10.1007/978-3-030-03023-0_10

Pipalia, K., Bhadja, R., & Shukla, M. (2020, December). Comparative analysis of different transformer based architectures used in sentiment analysis. In 2020 9th International Conference System Modeling and Advancement in Research Trends (SMART) (pp. 411-415). IEEE. https://doi.org/10.1109/SMART50582.2020.9337081

Saini, M., Adebayo, S. O., Singh, H., Singh, H., & Sharma, S. (2023a). Sustainable development goals for gender equality: Extracting associations among the indicators of SDG 5 using numerical association rule mining. Journal of Intelligent & Fuzzy Systems , (Preprint), 1-12. https://doi.org/10.3233/JIFS-222384

Saini, M., Sengupta, E., Singh, M., Singh, H., & Singh, J. (2023b). Sustainable Development Goal for Quality Education (SDG 4): A study on SDG 4 to extract the pattern of association among the indicators of SDG 4 employing a genetic algorithm. Education and Information Technologies, 28 (2), 2031–2069. https://doi.org/10.1007/s10639-022-11265-4

Salkind, N. J. (2010). Encyclopedia of Research Design. Encyclopedia of Research Design. https://doi.org/10.4135/9781412961288.n100 .

Salleb-Aouissi, A., Vrain, C., Nortet, C., Kong, X., Rathod, V., & Cassard, D. (2013). QuantMiner for mining quantitative association rules. The Journal of Machine Learning Research, 14 (1), 3153–3157.

Setiawan, R., Budiharto, W., Kartowisastro, I. H., & Prabowo, H. (2020). Finding a model through a latent semantic approach to reveal the topic of discussion in the discussion forum. Education and Information Technologies, 25 (1), 31–50. https://doi.org/10.1007/s10639-019-09901-7

Shah, M., Shenoy, R., & Shankarmani, R. (2021). Natural language to Python source code using transformers. In 2021 International Conference on Intelligent Technologies (CONIT) (pp. 1-4). IEEE. https://doi.org/10.1109/CONIT51480.2021.9498268

Singh, M., Saini, M., Adebayo, S. O., Singh, J., & Kaur, M. (2023). Comparative analysis of education policies: A study on analyzing the evolutionary changes and technical advancement in the education system. Education and Information Technologies, 28 (6), 7461–7486. https://doi.org/10.1007/s10639-022-11494-7

Sivakumar, S., Venkataraman, S., & Selvaraj, R. (2016). Predictive modeling of student dropout indicators in educational data mining using improved decision tree. Indian Journal of Science and Technology, 9 (4), 1–5. https://doi.org/10.17485/ijst/2016/v9i4/87032

Sverdlik, A., Hall, N. C., McAlpine, L., & Hubbard, K. (2018). The PhD experience: A review of the factors influencing doctoral students’ completion, achievement, and well-being. International Journal of Doctoral Studies, 13 , 361–388. https://doi.org/10.28945/4113

Rastogi, R., & Shim, K. (2002). Mining optimized association rules with categorical and numeric attributes. IEEE Transactions on Knowledge and Data Engineering, 14 (1), 29–50. https://doi.org/10.1109/69.979971

Tinto, V. (2006). Research and practice of student retention: What next? Journal of college student retention: Research. Theory & Practice, 8 (1), 1–19. https://doi.org/10.2190/4YNU-4TMB-22DJ-AN4W

Vilser, M., Rauh, S., Mausz, I., & Frey, D. (2022). The Effort-Reward-Imbalance Among PhD Students–A Qualitative Study. International Journal of Doctoral Studies, 17 , 401–432. https://doi.org/10.28945/5020

Walker, G. E., Golde, C. M., Jones, L., Conklin Bueschel, A., & Hutchins, P. (2008). The formation of scholars . JosseyBass.

Wendler, C., Bridgeman, B., Markle, R., Cline, F., Bell, N., McAllister, P., & Kent, J. (2012). Pathways through graduate school and into Careers. Distributed by ERIC Clearinghouse.

Willging, P. A., & Johnson, S. D. (2009). Factors that influence students' decision to dropout of online courses. Journal of Asynchronous Learning Networks, 13 (3), 115–127. https://doi.org/10.24059/olj.v13i3.1659

Wollast, R., Boudrenghien, G., Van der Linden, N., Galand, B., Roland, N., Devos, C., De Clercq, M., Klein, O., Azzi, A., & Frenay, M. (2018). Who are the doctoral students who drop out? factors associated with the rate of doctoral degree completion in universities. International Journal of Higher Education, 7(4), 143. https://doi.org/10.5430/ijhe.v7n4p143

Yan, X., Guo, J., Lan, Y., & Cheng, X. (2013, May). A biterm topic model for short texts. In Proceedings of the 22nd international conference on World Wide Web (pp. 1445-1456). https://doi.org/10.1145/2488388.2488514

Yang, Z., Dai, Z., Yang, Y., Carbonell, J., Salakhutdinov, R. R., & Le, Q. V. (2019). Xlnet: Generalized autoregressive pretraining for language understanding. Advances in neural information processing systems , 32 .

Zaki, M. J. (1999). Parallel and distributed association mining: A survey. IEEE concurrency, 7 (4), 14–25. https://doi.org/10.1109/4434.806975

Zuur, A. F., Ieno, E. N., & Elphick, C. S. (2010). A protocol for data exploration to avoid common statistical problems. Methods in ecology and evolution, 1 (1), 3–14. https://doi.org/10.1111/j.2041-210X.2009.00001.x

Download references

Author information

Authors and affiliations.

Department of Computer Science, Guru Nanak Dev University, Amritsar, Punjab, India

Manevpreet Kaur & Madanjit Singh

Department of Computer Engineering and Technology, Guru Nanak Dev University, Amritsar, Punjab, India

Munish Saini

You can also search for this author in PubMed   Google Scholar

Contributions

All authors contributed to the data collection, analysis, and writing of the manuscript.

Corresponding author

Correspondence to Madanjit Singh .

Ethics declarations

Ethics approval.

The research meets all applicable standards concerning the ethics of experimentation and research integrity, and the following is being certified/declared true. As an expert scientist and along with co-authors of the concerned field, the paper has been submitted with full responsibility, following the due ethical procedure, and there is no duplicate publication, fraud, plagiarism, or concerns about animal or human experimentation.

Competing interests

It is to specifically state on behalf of all authors that “No Competing interests are at stake and there is No Conflict of Interest” with other people or organizations that could inappropriately influence or bias the content of the paper.

Additional information

Publisher’s note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Kaur, M., Singh, M. & Saini, M. Analyzing the relation among different factors leading to Ph.D. dropout using numerical association rule mining. Educ Inf Technol 29 , 375–399 (2024). https://doi.org/10.1007/s10639-023-12260-z

Download citation

Received : 11 July 2023

Accepted : 04 October 2023

Published : 07 November 2023

Issue Date : January 2024

DOI : https://doi.org/10.1007/s10639-023-12260-z

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Machine learning
• Qualitative content-based analysis
• Numerical association rule mining (NARM)
• Find a journal
• Publish with us
• Track your research

• International edition
• Australia edition
• Europe edition

Is a PhD the right option for you?

Too often starry-eyed students rush into a PhD without knowing what it entails or how useful it will be. Daniel K. Sokol discusses what you need to consider before taking the plunge.

• PhD or no PhD, explore a range of exciting careers on Guardian Jobs

Embarking on a PhD is a big decision. Not only will it consume three to five years of your life but, in some UK institutions, the failure rate exceeds 40%. During that time, the 'great work' (ie the thesis) will hover above the candidate like the sword of Damocles, even in moments of supposed rest. So when students say they are thinking of doing a PhD, I ask them why.

For most jobs, a PhD is unnecessary. I, and many of my PhD friends, dropped the title soon after our release into the real world. The initial buzz of having Dr before your name dims with time, and using the title in a non-academic context exudes more than a whiff of self-importance.

People also equate the prefix with a medical degree. On a plane back from Australia one year, I heard the call dreaded by doctors and title-wielding PhDs alike: "is there a doctor on the plane?" Sensing that my knowledge of grounded theory would do little to assist the feverish passenger, my wife, a medical doctor, volunteered to save the day.

If future income is a consideration, a PhD is worth little more than a master's. According to Bernard Casey, who published a study on the economic contribution of PhDs, male PhDs earn 26% more than those who could have gone to university but did not. However, men with a master's degree earn almost as much, with a 23% increase. For women, the difference is smaller still. Variations also exist within individual disciplines. Casey concludes: "PhDs in social sciences, languages and arts do not enhance earnings significantly for either sex."

When I enrolled on my PhD, I didn't care about so distant an issue as future income. Armed with three years of funding, I cared only about my subject and pushing the frontiers of knowledge, however modestly.

Enthusiasm fills the heart of most prospective PhD students, but this enthusiasm can soon fade. The drop-out rate for PhDs is high. In the United States, only 57% of PhD students obtained their PhD 10 years after enrollment. In the humanities, the figure dropped to 49%. In my department, four of us enrolled on the PhD programme in medical ethics; two completed it. Contrary to popular belief, a PhD is not intellectually difficult but it calls for discipline and stamina.

A PhD, especially in the humanities, is a lonely affair. Days are spent alone in front of a computer. Antidotes to the common ailments known as PhD fatigue and PhD blues are, first, choosing a subject that can sustain interest for several years. Often students realise after a few months that their topic is not as gripping as initially believed. An additional consideration, when selecting a topic, is whether the choice will bolster an academic career. Some topics lie on the fringes of the field and may raise eyebrows in reviewers of articles and conference abstracts and in interviews for lectureships. An obscure PhD is also poor preparation for teaching a broad curriculum to undergraduate students.

The second antidote is choosing good supervisors. Knowledge aside, a good supervisor should be willing to devote time to the thesis. Beware the elusive professor, however stellar his or her reputation. It is worth talking to a supervisor's past or current PhD students before making your request.

Sadly, stories of disastrous PhD experiences abound. Unsupportive or bullying supervisors, lack of institutional support, late or radical changes of topic, poor advice, unfair viva voce examinations – the list of potential woes is long. So common are such problems that, after representing an aggrieved PhD student at an appeals hearing, I founded a service to help university students appeal unfair decisions. A frequent fault of students is allowing problems to grow rather than nipping them at the bud; early intervention is key. When I ask eager students their reasons for enrolling in a PhD programme, I do not seek to dissuade them. My own PhD experience, and those of countless others, was positive. Meetings with my supervisors were regular and enjoyable. The viva (or oral examination), which lasted three hours, went smoothly. Although academic jobs were scarce, I was lucky to obtain a lectureship immediately after the PhD. My thesis may even have contributed, microscopically, to the field.

Too often, however, starry-eyed students rush into a PhD program with scant knowledge of what it entails or how useful it will be in the future. The drop-out rate would be reduced, and much misery avoided, if prospective students possessed a more balanced view of the challenges, as well as the joys, of the PhD.

Daniel K. Sokol PhD is honorary senior lecturer in medical ethics at Imperial College and director of Alpha Academic Appeals

This content is brought to you by Guardian Professional . To get more content and advice like this direct to your inbox, sign up for our weekly Careers update .

• Guardian Careers
• Career choices

Comments (…)

Most viewed.

01 Sep Graduate Student Graduation and Completion Rates – Long Overdue

We don’t have graduate student graduation and completion metrics

Why you ask? Great question! The data simply are not collected at a national level. And to be fair – states, higher education organizations, and some institutions know A LOT of information about their graduate students. They track them, run analyses on them — all with the aim of increasing graduate student completion, improving the graduate student experience, and becoming more efficient and effective.

But what about IPEDS?

They collect a ton of information. Indeed – IPEDS is my go-to resource for a lot of benchmarking (and I’m not just saying that because I’ve been an IPEDS Educator for over 10 years). IPEDS is THE most comprehensive national higher education data system in the world. (Yep – lots of institutions in other countries ask to participate – voluntarily.)

IPEDS has 13 integrated (that’s what the I stands for in IPEDS) surveys. [Survey sounds voluntary – so I call them reports because they are required if one wants access to Title IV funding and some other federal benefits.] Of the 13 IPEDS reports (list below), only one of the reports collects information about graduate student completion – the Completions survey. There are three IPEDS surveys that focus on undergraduate graduation rates (Graduation Rates, Graduation Rates 200 and Outcome Measures). Yet, none include graduate student graduation rates.

IPEDS Completions Survey

The IPEDS Completions survey collects data on the number of students that complete a degree in a 12-month period. Luckily – at the graduate level, the data are disaggregated by:

• Master’s degree
• Research/scholarship
• Professional practice
• Postbaccalaureate
• Post master’s

Additionally, the data are also disaggregated by:

• Classification of Instructional Programs (CIP) code
• Race/Ethnicity

(Of course, undergraduate completions are also collected, but since that’s not our focus in this article, we are just listing out the graduate information. You can find more details about the entire IPEDS completions survey in the IPEDS instructions .)

Here is a sample of what the collection screen looks like. This may give you a better of idea of the data collected. Note the column headers (CIP Code, Award Level, Gender, and Race/Ethnicity). IPEDS also collects information whether or not the program is a distance education program (bonus!)

Are there other Graduate Student Graduation and Completion Rates options?

NCES Sample Surveys

The National Center for Education Statistics (NCES) conducts a series of surveys by gathering information from a sample of students (aka ‘sample-surveys’). These surveys are sophisticated – ensuring a representative sample of students in the country . Because students are the focus of the surveys (not institutions), a small percentage of institutions are part of the collection. Again – the focus is on the students.

The collection of information is quite thorough:

• Data from the colleges and universities the student attended
• Data from financial aid records (if available)
• Phone or email questionnaire to the students

The downside to the thoroughness is that these surveys cannot be conducted on an annual basis. And – because most of them are longitudinal surveys, tracking students over upwards of a decade, the data take quite a while to collect. Some argue that information about a cohort of students that is six or 10 years old isn’t relevant to today’s decision makers.

Below is a list of NCES Sample Surveys that are focused on postsecondary education (aka – higher education). Of the five sample surveys, one (Baccalaureate and Beyond) had a graduate student focus with a 1993 cohort of bachelor’s students. The study tracked these students through 2003 – 19 years ago. And a few things have changed since then.

Just in case you want to know a bit about the findings on graduate students from that survey, Nevill and Chen (2007) wrote an article using the data. They found that: “Rates of persistence and completion were higher among students who entered graduate school immediately after earning a bachelor’s degree, who attended full time and enrolled continuously, and who enrolled in multiple graduate degree programs.”

But – unfortunately, the data are not structured in a way to tell us about specific institutions. So – good to have some information, but usability for decision-making diminishes when applied to an individual institution.

It is worth mentioning the National Postsecondary Student Aid Study (NPSAS), Graduate Survey – because that sure looks graduate focused – and it is. However, it isn’t focused on graduate student graduation and completion rates. Rather – you guessed it – the NPSAS survey focused on student aid. Below is a screenshot of the data categories (left hand side) that are available for the NPSAS, Graduate survey. Sadly, there are no graduate student graduation and completion rates information. [Sidenote – You too can explore any of the sample survey data in NCES’ awesome Datalab via Powerstats. I sooo wish this existed back in my dissertation days . Instead, I had to wait to gain access to the restricted components of the Beginning Postsecondary Study data, which elongated my dissertation time.]

Individual Institution Websites

However, absent a common standard, this means that institutions report in different ways, use different definitions, disaggregate differently, etc. In short, it is hard to compare graduate student graduation and completion rates from one institution to another. And – let’s assume that all institutions use IPEDS definitions for calculating undergraduate outcomes and simply apply them to graduate student cohorts – and then post the information on their websites.

One will need to painstakingly scour each institution’s website for the information – because each will house it in a different spot on their website. Then after finding it, a researcher/interested person would need to extract the information from the website or PDF and put it into a spreadsheet (or something that lends itself for analysis). The likelihood of someone doing this for over 2,000 graduate colleges and universities is 0 percent. [Yep – there are nearly over 2,000 colleges and universities that offer one or more graduate programs – so this impacts A LOT of institutions.]

This brings us full circle to our third and final option – and the teaser in the first line of this article.

National Student Clearinghouse (NSC)

The NSC is one of THE most well respected organizations when it comes to student level data. And why shouldn’t they be? They have been perfecting their work for nearly 30 years (started in 1993). The overwhelming majority of institutions report enrollment data to NSC because their system “talks to” the federal government’s financial aid reporting system (National Student Loan Data System – NSLDS). Once an institution provides their data for NSLDS to NSC, they get access to a lot of other NSC services. This data maven’s FAVORITE is StudentTracker . “StudentTracker® is the only nationwide source of college enrollment and degree data. Nearly 3,600 colleges and universities — enrolling over 99 percent of all students in public and private U.S. institutions — regularly provide enrollment and graduation data to the Clearinghouse.”

That’s just what we needed – right?

Yes – the institution collects graduate student outcomes data. But, the dataset is protected information because it is at the student level. (NSC takes the protection of student information VERY seriously).

So – the good news is, the data exist. In fact, the NSC has looked as some graduate student data in 2017. Further suggesting, NSC data can answer some very important questions about graduate student graduation and completion rates (more on that in a few paragraphs).

The bad news is that institutions can only query the NSC StudentTracker databases for their own students (and since the institution provided the data on completion – there is no point in asking NSC StudentTracker who graduated at one’s institution when your institutional research office has that info and reports it to IPEDS). This means that individual institutions can’t benchmark to other institutions without entering into a data/information sharing agreement with other institutions (a very long process).

Back to the good news — the data exist and institutions have been reporting these data to NSC for decades. Even better news – NO, zero, zip, zilch, nada – colleges and universities need additional reporting to answer important questions about graduate student success. Which has significant implications on ALL of higher education and our communities.

What questions can we answer with these data?

I’m glad you asked. I’m a classically trained researcher with a practical-based approach. So, while I do my due diligence in the research and statistics space, it also has to make practical sense too. That’s the lens that I use when attempting to answer the BIG question…

So what? So what if we know graduate student graduation and completion rates? What practical value does this add?

Here are just a few potential questions that we could answer with the data. Note – all of these can – and should – be sliced and diced by institutional type, size, student characteristics, etc. I come at this question from three different perspectives, based on my own academic and professional background.

From a policy perspective (My doctorate is in higher education policy – so this is the short list).

• To what degree do different financial aid programs contribute toward increasing graduate student success?
• Under what institutional conditions do minority graduate students increase their likelihood for completion?
• How do we leverage findings from increasing minority graduate student completion to have an impact on the number of minority faculty teaching at the undergraduate level? This might bolster undergraduate minority completion rates.

From a data perspective (I’ve been in the field of institutional research and have served as the head of IR offices for over 20 years).

• What is the time to degree by award level (e.g., master’s, doctoral research/scholarship vs doctoral professional practice)? And how does that compare across our peer and aspirational institutions?
• What are the trends of graduate graduation rates by peer and aspirational groups?

From an accreditation perspective (I’ve served as an accreditation liaison at multiple institutions, served on institutional accreditation teams, and been an accreditation reviewer.

• From which institutions can we glean graduate student graduation rate best practices? Sharing those practices with other institutions may help them improve their student successes.
• How can and should team reviewers evaluate graduate student graduation and completion rates such that it informs accrediting recommendations and actions?

Wrapping it up

I hope you are deeply encouraged by the fact that graduate student graduation and completion rate data exist – and are just waiting for us to discover it. In the meantime, while we all wait for national efforts to get underway, individual institutions can take action now by analyzing their existing data – and calculating graduate student retention and graduation rates — as well as thinking about better ways to measure student success.

Data source

The Survey of Earned Doctorates (SED) is the sole data source for Doctorate Recipients from U.S. Universities: 2019 . The principal elements of the 2019 SED data collection are described in the sections that follow. More detailed information and related technical tables are available at https://ncses.nsf.gov/sed/ .

Survey eligibility. The SED collects information on research doctorate recipients only. Research doctorates require the completion of a dissertation or equivalent project, are oriented toward preparing students to make original intellectual contributions in a field of study, and are not primarily intended for the practice of a profession. The 2019 SED recognized 18 distinct types of research doctorates. In 2019, 98% of research doctorate recipients earned the PhD.

Survey universe. The population eligible for the 2019 survey consisted of all individuals who received a research doctorate from an accredited U.S. academic institution in the 12-month period from 1 July 2018 to 30 June 2019. The total universe consisted of 55,703 persons in 448 institutions that conferred research doctorates in academic year 2019.

Data collection. Institutional coordinators at each doctorate awarding institution distributed the SED Web survey link (or paper survey form) to individuals receiving a research doctorate. Nonresponding graduates were contacted by e-mail, mail, or phone to request response to the survey. RTI International served as the 2019 SED data collection contractor on behalf of NCSES.

Survey response rates. In 2019, 92.1% of research doctorate recipients completed the survey. Limited records (field of study, doctoral institution, and sex) are constructed for nonrespondents from administrative records of the university—commencement programs, graduation lists, and other public records—and are included in the reported total of doctorate recipients. The survey response rates for 1980–2019 and the item response rates for 2010–19 are provided in the technical tables ( https://ncses.nsf.gov/pubs/nsf21308/ ).

Time series data changes. After a multiyear review of Doctor of Education (EdD) degree programs participating in the SED, 143 programs were reclassified from research doctorate to professional doctorate over the 2010–11 period. No additional reclassifications of EdD degree programs are planned. SED data are no longer being collected from graduates earning degrees from the reclassified EdD programs, and this has affected the reporting of the number of doctorates awarded by sex, citizenship, race, and ethnicity. Several figures in this report show the impact of the decline in number of doctoral degrees awarded in education from 2009 to 2011 (see figure 8 and figure 12 in the section “ Fields of study ,” and figure 22 in the section “ Postgraduation trends ”). Readers should note that the declines from 2009 to 2010 and from 2010 to 2011 are at least partly attributable to the EdD reclassification.

Data license. Microdata from the SED may be obtained through a restricted-use data license (see https://nsf.gov/statistics/license/index.cfm ).

All Departments: PhD Completion Rates Statistics

• Durham University ranked highest since 2020 in QS World University Rankings
• Self-nominations open for Durham SU’s first ever part-time Officer roles
• Plans for new student accommodation withdrawn after MP criticism
• Planning application submitted for redevelopment of “failing” Prince Bishops Place shopping centre
• Construction of New College Durham teaching centre begins at Langley Moor 

Durham's Official Student Newspaper

One in four PhD students drop out

By Luke Payne

One in four Durham PhD students leave study without achieving their target doctorate degree. The data, acquired through a Palatinate Freedom of Information request, reveals performance is particularly poor in certain departments, with Computer Science, Education, and History performing amongst the worst in their respective faculties.

Palatinate received data from 10 UK universities counting the number of PhD students who left study between the 2015/16 and 2019/20 academic years and how many of those students received a doctoral degree. The figures include some medical students and other doctoral awards, while a small minority of students leave study due to transferring to another university.

At Durham, only 74% of doctoral students, who left study over the five-year period, received doctorates. This figure compares poorly with rival institutions such as Glasgow and Imperial where figures are 93% and 91%. Only two of the 10 institutions surveyed had poorer PhD success rates based on the raw data. These were Strathclyde (68%) and Liverpool (65%).

In response to the article, Durham University disputed Palatinate’s findings and said the true success rate was closer to 81%, after taking into account students transferring between institutions. It further claimed that it was not accurate to compare its data with other UK universities.

In the Faculty of Science, a number of Durham departments are very successful at ensuring their PhD students graduate with doctoral degrees. Over 90% of Physics, Mathematics and Earth Sciences PhD students graduate with doctorates.

By comparison, only 73% of Engineering and 69% of Computer Science students leave with doctorates. These rates are 10 to 20 percentage points behind comparable departments at Glasgow, Imperial, York and Leeds.

A Durham Physics PhD student told Palatinate that their department had a “very strong sense of community and provided many activities, both academic and social. This really helps students to feel less isolated, which I think can be a problem in other departments.”

“Only 55% of Durham’s Education PhD students graduate with doctorates”

In the Faculty of Social Sciences and Health, less than two-thirds of Durham PhD students within the Sociology Department, School of Government and International Relations and the School of Education graduate with doctorates.

Only 55% of Durham’s Education PhD students graduate with doctorates compared with 86% at the University of Leeds and 81% at the University of Glasgow. This equated to 51 Durham Education PhD students leaving with non-doctoral degrees (such as Research Masters) and 36 leaving with no degree over the past five years.

Within Arts and Humanities, the success of PhD students studying within the Theology & Religion, History and Classics and Ancient History departments was less than 73%.

In their response to the article, Durham University provided some explanation of why some departments’ PhD students may be struggling to achieve doctoral degrees.

“A number of our disciplines (Education, Theology and Religion and Business/Economics/Management) have a significant proportion of mature students returning to higher education… financial and life challenges and changes for these students tend to contribute to withdrawals.

“Science research is often conducted in teams whereas, in other disciplines, postgraduate researchers tend to work alone on their individual research topic. Differences in the proportion of part-time, mature, international and self-funded students between disciplines also contribute to differing outcomes.

“The University, through its Research Degrees Committee, analyses thesis submission and withdrawal data annually. Where a department has a submission rate that is lower… departments are required to review the data and report on their action plans.

“Recently departments have addressed this issue through a number of measures across the student journey that have shown positive results in improving submission rates. In particular, departments have reviewed and enhanced their recruitment practices and processes to support students through annual progress reviews and training.”

“It makes me concerned that the department really does not value its postgraduate researchers”

Not all students are satisfied that sufficient progress is being made to support PhD students during their programmes. Upon learning of Palatinate’s findings, a Durham Engineering PhD student provided this comment:

“It was disappointing to hear that only 73% of Engineering PhD students receive a doctorate. It makes me concerned that the department really does not value its postgraduate researchers, especially after very little support has been provided during virtual working since the start of the pandemic.

“Many PhD colleagues in Engineering haven’t had any additional support or check-ins from supervisors – if anything, it has subsided. It makes me wonder if the general lack of any community amongst Engineering PhD students and staff contributes to the higher drop-out rates.

“I envy other departments which have reading groups, specialised research communities and department organised socials. This lack of support has made me feel lonely, frustrated and underappreciated for the research and teaching contributions my PhD colleagues and I make.”

Image:  Michal Jarmoluk  via  Pixabay

• ← Another letter to my past self
• £1m+ individual donors to Durham revealed →

More by this author

£700,000 spent on Vice-Chancellor’s residence refurbishment

Archaeologists biggest spenders in staff expenses figures

Durham University to trial ‘hybrid working’ next academic year

All aboard(ed up): Durham’s former train stations

Durham PhD theses sold on Amazon without authors’ consent

The Legacy of Durham’s 900-year-old Hospital

Donate to palatinate.

Welcome to Palatinate’s website. We rely on readers’ generosity to keep producing award-winning journalism and provide media training opportunities to our team of more than 100 students. Palatinate has produced some of the biggest names in British media, including Jeremy Vine, Sir Harold Evans and George Alagiah. Every contribution to Palatinate is an investment into the future of journalism.

School of Business graduate program surges in top rankings

A U.S. News & World Report methodology change rewards schools for job placement, graduate earnings

The UC Riverside School of Business increased 20 positions in a U.S. News & World Report 2024 graduate school ranking released this week. UCR’s business school was the benefactor of a methodology change this year that places greater emphasis on post-graduation earnings and how quickly graduates found jobs.

The School of Business’ Master of Business Administration program ranked No. 90 on the business school list, released Tuesday, April 9. Though 506 accredited institutions were canvassed, only 124 U.S. graduate-level business programs provided sufficient information on post-graduation earnings and job placements to merit inclusion on the U.S. News list.

“The school’s ascent in the rankings is a testament to its strategic initiatives, such as the expansion of its facilities and the enhancement of its academic offerings, which align with the evolving demands of the global business landscape,” said Rami Zwick, associate dean of graduate programs for the School of Business.

The rankings released this week are based in part on reputational surveys sent to more than 15,000 academics and industry professionals, including corporate recruiters. In the surveys, deans, program directors, and senior faculty are asked to judge the academic quality of a program.

In addition to the surveys, scoring factors selectivity — gauged by graduate exam test scores; undergraduate GPA, and acceptance rate. About half of the score is based on a change this year that places greater emphasis on earnings — assessing post-graduate salaries by profession — and successful job placement. The revised job placement metric rewards business schools when their graduates get jobs quickly — either when they graduate or within three months of graduation.

Separately, the business school’s part-time MBA program was ranked No. 73, an increase of three positions from the past year. That ranking was among 269 universities.

In fall 2024, the School of Business will mark the 55th anniversary of its founding and the 30th anniversary of the A. Gary Anderson School of Management with the opening of a new 63,400-square-foot academic building .

UCR’s School of Education was ranked No. 86 out of 237 positions in the graduate school rankings, which considered only doctoral-level education programs. Last year, the school ranked No. 80.

The education rank is based on research expenditures; assessments by education-school and graduate-school deans and by professionals including recruiters and school superintendents; total degrees awarded; student-faculty ratio; faculty awards; and selectivity, measured by acceptance rates.

U.S. News this week postponed publication of its rankings for medical and graduate engineering programs after some universities questioned the rankings methodology for those programs.

U.S. News & World Report, the standard-bearer among college rankings, publishes its anticipated undergraduate rankings every fall. In the most recent rankings, released in fall 2023, UCR climbed 13 positions to No. 76 overall among U.S. private and public universities. UCR was ranked No. 2 nationally in social mobility, which considers the degree to which a university elevates its graduates to a higher standard of living. 

Media Contacts

Related articles.

Have career goals. Will travel.

Creating Digital Equity for UCR Students

Linking Generations of Highlanders

$5 million grant supports community service by UCR students