Why You Should Quit Your PhD

A long time ago, I had a very serious discussion with my PhD supervisor. I was depressed, I felt inadequate to the task of completing my PhD and I really just wanted out.

Actually, that’s not entirely true. I wanted the PhD but I could no longer see the path to finishing it. So, I spoke the words that no supervisor wants to hear: “I’m thinking of quitting”.

His response really surprised me. He looked at me with a smile and said, “That’s great — 1 out of 3 is out of the way.” I had no idea what he was talking about. He explained that in his experience, a student will seriously consider leaving their PhD at least 3 times. This was my first time and we just needed to work through it.

I still felt like shit. However, when I think back to this moment I realise he gave me something very important. He normalised the idea of quitting for me and made me feel a tad less alone than I had before. I felt more supported.

I still wanted to quit.

The expectations suck

A PhD can feel like a jail term. You cannot just up and leave part way through your sentence. But why the hell not? What if a PhD is not for you? What if the vision for your future they sold you is not what you really desire? Why is a PhD treated so differently to other jobs?

I believe a big part of this is that your entire education has been building to this point — it’s been a rather straight line. One piece comes after another and if you want to be a professional in your field then this is the next step on the path. If you quit your PhD then the entire journey would come crashing down. What on earth would you do? The singular goal would be lost. Your chance of getting a job would be sweet FA.

Of course this is a line of bullshit that we have all been fed to keep us in our PhDs. In reality, by the time you have reached this level, you will have gained an enormous range of skills that are applicable to many different industries. Therefore, the idea that it has all been wasted if you quit, just does not hold water. If you have reached the level of PhD student you are already in the top few percent performance wise so will likely be highly adaptive to new types of work. The only real waste is the project work that you will have started in your PhD, but if you are going to change fields then why would you care?

Keep in mind that PhD students are the engine room for universities and research institutions. The drive to keep you in the house will be very strong. You are free and very skilled labour. If you finish there is an added cash incentive for institutions. It’s a win win for the institutions if you complete. To be fair, it costs an institution a lot of money for you to have the facilities to do your PhD, so completing on time allows them to get some of that expense back. It’s odd that most institutions ignore the mental health of PhD students given how much is at stake.

Unfortunately as a society we don’t match our training numbers with what the workforce actually needs — so in essence a few people quitting wouldn’t be a bad thing. You would be doing your colleagues a favour if you quit, as they would have less job competition when they finish.

Have I convinced you to quit yet?

Well let’s talk about what science is all about. That should do the trick.

Apologies if you are in another field, but it’s not difficult to draw parallels. When you enter a PhD you are told about the interesting areas you will get to explore. You will be told how you will push the boundaries of human knowledge and how fun and interesting this will be. You will get to publish, discover new things, travel the world……and the list goes on. Sounds bloody great. But hang on, did anyone actually tell you about how science actually progresses most of the time? I’m guessing not or you would not be reading this article.

In many ways, science is about failing. Unless you are doing incremental work, which in itself can be boring and only mildly useful, the majority of science involves failing a lot. If it was easy it would all have been done by now. If it was easy you would not be getting a PhD for doing the work. So if you are the sort of person who does not like to fail — and lets be honest, what kind of idiot actually likes to fail — then perhaps a PhD is not for you.

I think there are numerous reasons why you might be thinking about quitting. You might be in that rut where nothing new is coming out of your work — the classic scenario where your effort/reward ratio is seriously out of whack. You might be finding that your work is unfulfilling and you are struggling to find meaning. In reality, this is probably reasonable because early in your career the chances of your work being impactful are rather low. You might feel like an imposter (that’s another article I need to write!).

Perhaps you have an unsupportive supervisor. This can take many shapes and sizes. Some supervisors push too hard. Some not hard enough. Some try to clone themselves disregarding that different people work and are motivated in different ways. Some are just A-holes in general, and some lack the field expertise and innovative thought to provide inspirational projects to students. A supervisor can make or break a student regardless of the standard of the student. If you have one of these supervisors then you should give solid thought to getting the hell out of Dodge.

Before you quit…

I would like to propose that a PhD is not that much different to other jobs. It’s something that, if you cast off the centuries of cultural heritage from universities, should be seen as something to try and walk away from if it is not right. A PhD should never come at the cost of your mental health.Institutions almost never have systems in place to monitor and address the health of PhD students.

If at this point you have already decided you are definitely going to quit then don’t read on. If on the other hand you have not quite steeled yourself to withdraw, then here are a few words to help you endure what is still ahead of you.

Firstly, I suggest that you reset in your mind about what science is really about. A real scientist is not someone who wins every day in their work. It is someone who has the ability to notice something different when it comes along and recognise that difference for its potential. Sometimes we can wait 12 months or longer for this one day to occur. We need to be ready to grab hold of it when it happens. That’s what science is about, it’s not about success every single day and the best stuff is not about incrementalism. It’s easy to lose sight of this distinction especially in a system where innovative and risky thought is discouraged by funding agencies. Science is as much about eliminating the things that don’t work, as it is finding the ones that do. That’s not failure, that’s progress.

There will be times when it feels like certain things just won’t ever move forward. This is what discussions with your supervisor are for. They have more experience and given the opportunity, they should always be able to suggest a course of action. Sometimes that action will be to abandon the approach and move onto something else. These decisions should feel like a partnership between the supervisor and student. If they don’t then the supervisor is not doing their job.

The relationships you have in your PhD are very important. I encourage people to develop three types. The first is the supervisor relationship. This person is not your mentor. A mentor must have no personal gain involved in your work if they are to give you advice. The supervisor by contrast is there to guide your research journey and provide some career input. They need to know how you work best, how you like to get feedback, what sort of work you enjoy, what your weaknesses are. The more you give them the better they can help you navigate the difficult research path. Of course, you need to know the answers to these questions as well.

The second type of relationship that must be cultivated is the connection with others within your research group — especially those doing PhDs and postdocs who have recently completed. These are your support people and generally will all have experienced difficulties themselves. If they have not, they will at some stage, and the more open you are about the challenges the more these people can provide support. However, be clear about the sort of support you need.

I recall the group meetings we had when I was a student. We rarely talked about successes. We mainly talked about problems we were having and the group would rally to help solve them for the individual. It felt incredibly supportive and it prevented things from festering.

The third support group you will need is people from outside your discipline, even potentially outside your institution, but still doing PhDs themselves. People who can understand the challenges but are not necessarily within the political bubble of your lab.

Ultimately though a PhD can feel like a lonely enterprise. So what else can you do? For a start, it is important to keep a list of all the compliments you have received over the course of your time in the job. Whether written or oral, it can be very valuable to have these together in one place so you can reflect on them when feeling beaten. It doesn’t matter where they came from or how you keep them. Be kind to yourself and acknowledge that you are certainly capable even if the work is challenging. Don’t associate failure in the work with your failure — some research paths will simply not work.

No matter where you are in your PhD you will have made some progress. It’s important to remind yourself of where you were and what you knew 3 months ago, 6 months ago, a year ago. We all progress but we forget to give ourselves credit for that when things get tough. Typically, I find that people are closer to the end than the beginning, but the end seems far away.

What we don’t get taught prior to a PhD is that we need to learn to be exceptional project and time managers. A 3 to 4 year project is immense and it takes a lot of detailed planning. Most PhDs don’t have this and they certainly don’t have contingency plans when things don’t work out. Open discussion about these issues with supervisors and peers is necessary for all projects to be successful.

Most important though is maintenance of your health during a PhD. On occasion, this will take a team of GPs, psychologists, friends and family. Take prevention and treatment seriously. Compromised health reduces your ability to think effectively and you will see everything about your PhD progress through dark clouds. Always remember that the cost of a PhD must never be your health. Institutions will rarely have health support programs in place specifically for PhD students so you will need to address this yourself.

So before you quit, be sure to strip away the institutional attitude and pressures around quitting so you can make an informed and personal decision. A successful PhD depends on health, training, commitment, good supervision and supporting infrastructure. If any of these are lacking, the entire endeavour is unnecessarily harder.

Do not quit for these reasons:

1. Your experiments aren’t working

2. Your supervisor is an arsehole

3. Your supervisor is incompetent

4. You are finding the work too hard

5. You feel like an imposter

6. You don’t have the tools you need

7. You feel alone in your work

8. Everyone around you is having wins and you are not

9. You had some negative feedback

10. You are struggling to write

11. You have no publications

Consider quitting for these reasons:

1. You are trading your mental health to remain in the program (health must come first!)

2. You have learned what research is about and you no longer find it interesting or right for you

3. The longer you stay, the more you look towards other more interesting career paths

So please go ahead and quit your PhD, but make sure you are doing so for the right reasons so you can live with your decision 10 years from now. Don’t let others take this from you — you have worked too hard to get this far. If you leave, it needs to be a positive move, not a retreat.

Shane HuntingtonAugust 1, 2019

How to Improve Australian Research Grant Systems & Support Researcher Mental Health at the Same Time

A few months ago, I joined a friend for a coffee and we started talking about the NHMRC grant system. The moment it came up, I noticed her demeanor shift. She seemed beaten. She didn’t even know if she had won or lost the grant. The process alone was enough to damage her confidence. I had supported her in the grant writing process so I knew it was an excellent submission. But in the current system, quality doesn’t seem to matter, and everyone knows it.

So I asked her, would she be willing to trade the current system for something else? I proposed a short application of two pages, with a 70% cull of applications followed by a simple lottery. “Absolutely”, she responded. This would be substantially better than the current system. Then she got excited because she realized the lottery part could be electronically resolved in a day.

Researchers would not be concerned about their careers right up until weeks before contracts are about to end. We would not pretend that quality was the main factor in making decisions? We could have a system that was not biased against women. We could make sure that old white men with large research teams didn’t get most of the cash. We could guarantee that truly innovative ideas get funded even if they are risky. We could stop making 90% of the research community feel like garbage.

Currently, we have a system that is taking between 3–6 months of a researcher’s time to be part of, another month or so if you are a reviewer, and we could trade that for something that would probably take a few days. Importantly, researchers would actually see this alternative as fairer.

Make no mistake, the research grant systems in Australia are doing real damage to mental health. When the system itself causes damage to the very people it is supposed to support, something has to change. The current situation is not okay.

The lack of care for mental health goes well beyond the research bodies themselves.

We put the care of our researchers at the very bottom of the pile across the sector. It’s institutionalized. There is a lot of talk of mental health, but institutions do nothing to rectify these top issues. Show me a researcher that is happy with the current grant systems and I will show you 100 that are not.

There is little doubt that having more money in the system would alleviate some of the problems. Money, however, is not the only issue that exists. If we really want to address problems with our granting systems, we need to look deeper. We need to understand just how much damage a bad system can do to the mental health of our researchers.

The existing systems are so bad that even when researchers are successful they tend to keep quiet about it because they know so many around them are suffering. They have “grant success guilt”.

The Top Issues with the Current Grant Systems in Australia

Before running through this list, I want to be clear that I see tertiary and independent research institutions along with government agencies as part of the “system”. It’s not just the NHMRC or ARC that cause all the problems. It’s a solid team effort.

This list is not exhaustive and is in no particular order of priority, but obviously #1 is the most important. Not all problems apply to all granting agencies.

#1 — Time wasted: Months of research time on long applications

#2 — Waiting for results: Unknown outcome dates, year after year

#3 — Timing of applications: Discriminatory for parents

#4 — Reviewer feedback: No quality control, sometimes hurtful

#5 — Impact of grants on overall workloads: Extreme, up to 30% of the year (or more?)

#6 — System biases: Race, gender, age, field

#7 — Success rates: Devastatingly low — career ending

#8 — Institutional pressure: Extreme pressure from institutions to apply

#9 — Removal of review panels: No oversight of reviewers

#10 –Track record assessment: Field biased, not comprehensive

#11 — The reviewer lottery: Vast standard differences

#12 — A system that is not adaptive: Difficult to improve grants based on feedback

#13 — Decision making transparency: Completely opaque system

#14 — Relative to opportunity: No real structure for assessment

#15 — ROI for universities: Salary expenditure on applications enormous

#16 — The Medical Research Future Fund (MRFF): Not viewed as a competitive system

#17 — Partial funding of grant budgets: Budgets arbitrarily cut without cause

#18 — Assessment based on academic level, not years of service: perverse incentive to stay junior

#19 — Preprints not allowed: Systems out of date with current publication practice

* * *

There is a consistent theme in almost all of the problems that are listed above. Above all, most of the issues are due to a lack of understanding of what is being communicated to researchers and what the impact of that will be. The proposed solutions below, are based primarily on a more nuanced communication strategy.

A Researcher-Centred Funding System

In healthcare, there has long been the goal of achieving what is often called ‘patient centred care’. For the majority of health systems this is still a goal, yet to be reached, but we all understand the need and the positive impact on lives and economies that would result.

Our grant systems need to follow a similar path. The public investment in research needs to get to our researchers in a way that is positive, competitive and efficient. Architecturally, the wellbeing of our researchers should be at the very centre of any design and there should be strong support for the system even by those who fail to be funded. So basically the opposite of what we have right now.

Despite recent reviews and redesigns of some of our granting systems, many of the most significant issues still exist. To be blunt, any system that requires more than a 20-page application for $100k in funding is simply not fit for purpose. Any system that discriminates based on family commitments, gender, location, background or seniority is not fit for purpose.

What I am proposing below is not a complete redesign of the system, (that will take a lot more time than I have in this article). Instead, I am suggesting a few rapid changes that hopefully will take care of the majority of the big issues that have been listed. This is not an exhaustive list, nor will it address all of the issues raised.

Change #1 — Expression of Interest Model

Over the last 25 years, I have engaged with a lot of grant systems. I’ve been an applicant, a reviewer and a consultant. The most effective schemes I have come across all have a two tiered application process. Typically, this requires an initial expression of interest submission that would be short and sharp, followed by a small selection of projects being invited to submit a more substantial application.

The expression of interest application must have the goal of rapid communication with minimal time spent. Researchers will be aware that some 70–80 percent of these applications will be rejected. The full length of these applications should not exceed 10 pages and preferably should be limited to 5 pages. With much shorter applications, more reviews can be done for each applicant. The goal, given these are about 1/10 of the length of the current applications, should be for a minimum of 10 reviews. If you are going to cull 70–80 percent of people, you had better have some good statistics to back you up — more reviews will help with that. Limits on the number of applications would still need to apply.

Every review must outline 5–10 ways in which the expression of interest could be improved. Preferably this feedback would be according to deliberate instructions from the funding agency. It should not be a free for all — applicants should have a fair idea of what areas they will get help with. Applications should have a simple budget during the EOI phase — in fact budgets should not be a factor for transition to the next round. The results of an EOI round should be delivered within 6–8 weeks.

The 20–30 percent of applications that make it through the EOI round would be asked to provide ‘complete cases’ for funding. To be clear though, even these should not be of similar length to those currently being used. These should be limited to 30–40 pages. Similar requirements for feedback would be required by reviewers. All final decisions would be made by expert panels.

Back in 2005 I applied for funding under the State Government of Victoria’s Science, Technology and Innovation program. After an initial 5-page application that resulted in a cull from 157 down to 23, we were all put into a workshop with a professor from the Melbourne Business School. He provided us with guidance on what a good business plan for government looked like. We all learned. We all felt like we were working with the funding agency. At no point did we feel our time was being wasted. Above all, our projects ended up being better designed because of the grant process itself.

Change #2 — Multiple Application Times

One of the benefits of reducing both applicant and reviewer time loads with EOI models is that you can have multiple grant rounds without increased burdens. Currently, there is significant discrimination in our grant systems due to applications being required proximate to school holidays and other significant dates. This can be reduced by offering multiple rounds per year — each of equal value. Ideally 3 rounds would be offered but with a rule that you could not apply in consecutive rounds. If only 2 rounds were offered, then this rule should not apply.

The outcomes of each round should be delivered before the next round opens. Having multiple rounds will also spread the ministerial responsibilities out which hopefully would lead to faster and consistent turn around. Having numerous rounds would also enable researchers to engage with the research grants systems shortly after they start new positions. Currently some staff on 1–2 year contracts that start just after application are due, have to wait almost a year to get their 10% chance of success. Multiple rounds would remove this delay and assist new appointments.

Change #3 — Decouple Ideas and Track Records

In the current system, researchers are chained to the past by the assessment of their track records. Ultimately a system that decouples the assessment of ideas from the assessment of track records would be more effective. Frankly I don’t care where a good idea comes from. It should receive top marks if it is well constructed and argued. The removal of bias should always be the goal and decoupling ideas and people will help at least partially achieve this.

Applications could be readily split and sent to different reviewer groups. The assessments would later be recombined and if there was a need to connect the two areas of assessment this could be done via expert panel discussions.

Change #4 — Relative to Opportunity Calculator

I believe it is time we started working on a system that would calculate a ‘relative to opportunity’ index that would be utilized in the review process. The bias in the system currently means that researchers have little faith in how significant life and career events are taken into account when their track record is being assessed. An automated system would not be perfect at this but I suspect it would do a far better job than our current ‘reviewer lottery’ system and it would level the playing ground for assessment. It would need to account for different fields and different types of employees — e.g. clinicians vs academic researchers etc. Careful design would be required here to guarantee that the system itself did not perpetuate further biases.

Change #5 — Use of Budget Bands

It would appear that almost nobody gets the budget that they ask for from grants. In the case of ARC Discovery programs, the average is about 70%. I don’t see the point of researchers going to great effort to demonstrate specific costing when that is ultimately ignored. I have never understood how institutions put forward proposals and then get told they have to do the exact same job with no negotiation for 70% of the cost.

With this in mind, we could save a lot of time by asking researchers to simply nominate a ‘band’ for funding during the EOI phase. The band ranges would need to be in increments for 50k-100k. More detailed budgets could then be required for the more detailed 2nd stage applications — but even then, if the 70% funding model continues the value of budget specificity cannot be argued.

Change #6 — Feedback Systems — Review the Reviewers Program

To solve some of the issues with reviews that are either unprofessional, lacking field expertise, or potentially biased, we should establish a ‘review the reviewers’ program. In this program every review generated would need to be read by one other reviewer. If there is a concern with the review, then the second reviewer would flag the review and it would then need approval by a 3rd party (possibly even an expert panel) before being sent to the applicant. This would be a simple pass fail program — if the review was deemed inappropriate it would simply be removed.

The act of having this program in place will in itself improve the standard of reviews and reduce the amount of damage that reviews can do to research self-esteem, mental health and careers. If the time load here becomes problematic, then it could be done on a percentage of reviews each year randomly. Not a perfect system but works for the tax office!

Change #7 — Transparency and Communication Charter

All grant systems should have charters in place that guarantee that there is transparency regarding decision making. Despite the elaborate systems in place right now there is a wide spread concern regarding fairness. Part of this no doubt comes from the crushing success rates coupled with extraordinary workloads for grant applications. But the lion’s share comes from the opaque nature of the review process and the inability of applications to iterate the standard of their applications from year to year.

Similarly, the effectiveness of communication in the system needs to be actively monitored. In situations where 100 page applications are required there has been a communication failure. Not only is such length unnecessary, it also ends up being ineffective because reviewers cannot effectively engage with the weight of the material. Every aspect of the process should be optimised from a communication perspective. Minimise time spent, maximize information quality delivered, at every level of the process.

* * *

It is the role of our funding agencies to distribute the public investment in research in a competitive, responsible and fair way. The current lack of feedback and empathy in the grant systems means they are failing in this task. Systems that do not have the support of their constituents need to be carefully modified.

There are many barriers to changing these systems. Some of these are significant but given the current negative impact on researcher mental health, I believe the effort is well and truly warranted.

Blog

Why You Should Quit Your PhD