Why don't PhD students work in pairs?

Question

I will speak from my experience as an integrated-Masters graduate in Physics in UK, but I think this is applicable to any discipline.

Universities got it right when it comes to laboratory experiments or research projects during the degree: they make students work in pairs. I loved my final year research project because being able to brainstorm and bounce ideas back-and-forth with my MPhys partner took away most of the daunting nature of research. Having someone right next to you working on exactly the same project can prevent students from feeling deeply disoriented and aimless. You support each other when you're feeling lost, and given you're both on the same boat, it takes away the natural response of procrastinating and losing motivation that comes with being alone in such situations. Thinking about things together as loosely as needed can give you a natural sense of motivation somehow and allows you to be a lot more productive and effective.

• More productive because together you are able to come up with a lot more ideas, and you motivate each other to keep thinking for longer periods of time without giving up.
• More effective because the ideas coming from you and from your partner build on top of each other, and gain strength to penetrate into the unknown. When you are on your own, it's more likely that your ideas will stack up at the bottom, not having the same penetration potential.

A study in the UK reported that 46% of researchers feel lonely at work, which increases to 64% for PhD students. I believe the loneliness of independent research is the real root of the problem with Academia, the reason behind the mental-health isues such as impostor syndrome, anxiety and depression.

I looked up on Google 'Why don't PhD students work in pairs' and there are no results at all. Why have we normalised that research should be done alone? Yes, you have your supervisor to discuss things with maybe more than once a week if you are lucky, and yes, you have other PhD students around you to socialise and discuss ideas with, but this is not the type of loneliness I am referring to. It is the loneliness of independent research, where you get to your 3rd, 4th year of PhD, and being so advanced in your project means nobody understands the nits-and-grits of it like you do and hence no one can really help you anymore.

---

Now, I can think of a couple of issues with pairing PhDs up:

1. Money. Universities don't want to pay twice the money for a single project. Say you have the same number of PhD students, but they're all paired up. The pyramidal scheme of Academia wouldn't want to halve the number of projects for the same amount of money. That is less publication potential so nah. They are increasingly putting up more Mental Health support services, but I think they are not tackling the root of the cause, just the symptoms.

However, the fact that putting students in pairs could make them a lot happier, more productive and effective, means this could be turned around and it could actually end up being benefitial for the research output.

2. How do you match pairs up? Pairings students up randomly could go wrong if they don't end up getting along well. And some students might even prefer just working on their PhD on their own.

Just make it a normal part of a PhD offer to let the student decide if they want to undertake their PhD alone or with a partner. And put some formal procedure in place in case a pair of students really did not manage to get along. The procedure could just involve splitting the directions of research of the pair so they can carry on alone, or join someone else's project (as long as it's sufficiently related) who might have been split from their previous partner but would still want to work in pairs.

3. PhD is preparation for independent research. PhDs who manage to finish the programme on their own show they are prepared to be successfull in climbing up in Academia, in case they choose to apply for a Postdoctoral research assistant job.

If PhDs could hugely benefit from working in pairs, the same idea could be extended to Postdocs. I understand once you get to positions higher up in Academia, especially after tenure, one gets so many responsibilities (meetings, lectures, tutorials, supervisions, etc) that having a pair to do your research with becomes unfeasable. But still, given you have a fixed office in a department, nothing prevents you from having a collaborator who you work with as closely as you want.

---

Are there any other reasons I have missed which makes this unfeasable? Am I think about it wrong? If University knows pairing students up during their undergraduate is benefitial, why discontinue that afterwards?

Answer 1

In the labs I have worked in, graduate students discuss their projects with everyone else in the lab: other grad students, post docs, technicians and other specialist staff, undergraduate students, their advisor, other collaborating professors and people from their labs. They do this at every stage of their project: not just years 1 and 2.

The lab I am in right now currently has just one graduate student; she works closely with our lab manager who has over a decade of experience doing the sorts of experiments required for her thesis project, and she co-supervises three undergraduate students. She also has a co-supervisor who has two other graduate students and the three of them work in tandem, though they all have their own personal niches and projects. If you had her working in a "pair" instead of this whole research community, she'd be losing a ton of interactions with others, not gaining any.

Not all labs are like this of course, and I'm sure things vary a bit by field. Biology is a pretty communal research field compared to, say, pure mathematics. Working collaboratively isn't for everyone, and causes all sorts of problems: who does what, who gets the "credit", who makes decisions when there is disagreement. That isn't to say that collaboration isn't a good thing, but it seems like projects work best when there is someone who has ultimate responsibility for that project. In industry you will find people with a project manager role; in academia, there are no such project managers, so everyone needs to manage their own project to some extent. It's a skill that will be required eventually if a student is going to stay on the academic track; I don't see why they should wait until later to learn it rather than to start as a PhD student.

Answer 2

One particular issue of the proposal is with the matching up part. It can be hard enough to recruit one capable PhD student. Finding two capable PhD students with similar backgrounds and interests that can work together sounds like a nightmare. Sure you might succeed sometimes, but doing so consistently over all PhD hires sounds near impossible.

And even you have managed to match a pair at start of their PhD, that does not mean it will work out for 4 years. Since PhD projects are pursuing novel research, they evolve overtime. A large part of this evolution is based on the candidates interests and strengths (which should also be evolving over time). Since people are different, the evolution for a pair of PhD candidates is bound to diverge leading to problems. (Or more likely the evolution is going to be dominated be the needs of one of the pair, impairing the potential for growth in their partner).

This leads me to the interpersonal problems this model is likely to cause. Four years is a long time for two people to put up with each other (many marriages don't manage to last this long!). If you are essentially binding two people at the hip, many pairings will develop issues overtime. You will probably end up have to send many to couple's counseling.

One particular issue that might come up: What happens if one of a pair decides they want to quit after a year? Is their partner going to be left as a single? Is this person going to be guilted into staying in a position in which they are unhappy?

In short, I am not convinced that this proposal would actually lead to happier, more productive PhD students in general. It might work for some, but I will great misery for others. This is not say that there are not a lot of bad practices in supervising PhD students. I'm just not convinced that this proposal would be a solution.

Answer 3

This is an interesting thought experiment.

I think your last point is the main issue - a PhD is a training course in becoming an independent researcher. Sure, you could move the point at which you become independent up a few notches, but that just means you'll have to learn things you would otherwise have learned as a PhD student later.

You compare this to a final year project in an integrated Master's course (effectively an extended Bachelor's degree and taught in a similar way). That's fundamentally different from the way that a PhD or research Master's works - you aren't being taught about existing research by replicating it in controlled conditions, you're conducting wholly new research. That involves demonstrating that you - yes, you, not your research partner - can come up with, conduct and defend well-conducted research. How would you disentangle each partner's contributions to a thesis?

As you identify, money is important too - when you get to postdoc level and above you're spending large sums on doubling up each post. Staff costs are (at least in my field) the vast majority of grant costs so you'd have to drastically increase your grant income (this is more the problem than a university being cheap - you have to show that this is value for money for a funder).

Answer 4

This is from the standpoint of someone who has extensive experience and writings with "Agile Processes" including pairing in both computing and management. I've supervised a few doctoral students and would have been open to considering a "joint" project by a pair of students. I've thought about the issue, but it never actually happened. I've also done quite a lot of "pair teaching" in graduate programs.

However, pairing is a step beyond collaboration. In particular, it doesn't mean dividing up a problem into parts on which people work independently, even if they try to bring it together at the end. That "dividing" happens now in large scientific labs where an overall goal can be attacked with smaller, independent, goals, and doctoral students taking responsibility for one of them. The answer of Ian Sudbery explains this collaboration process very well.

In pair teaching, for example, both professors are present for all lectures and participate in all teaching activities. One of us would have the "floor" at a time while the other observed and made comments as necessary, perhaps also responding to some student questions independent of the "lecturer" of the day - though we didn't actually do a lot of what you can call "lecturing". But that is another matter. In any case, collaboration (team teaching) is much easier to arrange than pairing.

• Why it doesn't happen

One reason that it doesn't happen is that there is no tradition for it and, hence, some fear of the unknown. There is a rational basis for that, actually. As an advisor, I take on some responsibility for the future career of my students and I don't really have evidence that if I stray too far from traditional practice that it won't reflect badly on the students. I can't put them at risk.

Moreover, the two students in a pair are unlikely to be hired together to continue their work. I think that would be so rare as to be impossible to measure.

First, for it to truly be called "pairing" both students would need to be actively engaged in all practices all the time. Dividing up a project in to two projects isn't pairing, though it can be collaboration. I'll stick with that more strict concept. A "dissertation" with two authors would need to pass a quite strict filter. First, the university would need to agree that it is acceptable, and award two degrees for one dissertation. More important, there would need to be some evidence, hard to obtain, that the participants don't suffer in the academic marketplace.

I think the above reasons are far more important obstacles than money or the "matching" problem. The latter is trivial. Permit it only when two students request it themselves. If they come to me with a proposal, then I'll consider it (with suitable caveats and warnings, of course).

• How it might happen

If two students came to me (past tense, as I'm retired), with a proposal for a paired project toward their dissertation, I would consider both the students and the nature of the project. This wouldn't have been a big surprise, since I integrated this sort of thing into teaching and much of their other doctoral studies was done in a team environment. They were used to working with one another and sharing things.

I wouldn't worry too much about "equal work" since the environment let me keep a close eye on what was going on, and peer evaluation was used as appropriate (not peer grading). We had very good 24/7 communication facilities.

But the nature of the project would be key in my view. It isn't that the project would have "parts" but that it would have "aspects" that might develop in to distinct dissertations based on their paired work. Sorry, but I don't have a good explanation of "aspects" since I don't have an example in mind. But technical vs managerial aspects might have been able to pass the test. Or, perhaps theoretical vs applied.

And note, that I'd still be wanting to see two dissertations, probably cross cited in many places, but enough independence that the university doesn't need to be involved and potential employers wouldn't worry about hiring either of the students.

All of this suggests that it might be possible to manage this, but I think it would be rare in the short term. If some evidence of success would arise from it, then it might expand.

Answer 5

This is quite field specific. In biology it is very common for research to be a team undertaking. For example, the new project we are just starting has a team of 2 postdocs and 2 PhD students. Now, in general this isn't complete pairing, in that a given experiment will probably only be done by one person, and certainly one person at a time, but if things go well, this team will discuss things on at least a daily basis, if not more frequently.

Indeed, such is strength of this team attitude that one of the main roles of a thesis committee is to make sure a student has sufficient material they can call theirs that they can write a thesis that is based on their own work - in the end, in order to certify someone is a capable researcher, you need to know what they did, and what others did, and what were their ideas, and what someone elses. One cannot keep working with a pair forever - there is a reason that companies only have one CEO, or even one team leader.

There is also the credit issue - if things are done in a team, who comes first on the authorship list (as even where contributions are marked as "equal" or "co-first", people often distinguish between first-first and second-first).

I had assumed that this was also the case in physics, where paper authorships can run into the hundreds.

One distorted version of this is when two people put on the same project end up competing, rather than collaborating. This could be due to their personalities, but there are also stories of supervisors setting two people the same problem and saying "the first to solve it get the publication/grant". Giving people related projects, or specified parts of a whole, avoids this.

I have to say that far from loneliness, the team spirit and "lab families" I have encountered in science is one of the things I find most appealing about the job.

Answer 6

TL;DR: people are different and you cannot apply one mold to everyone.

Universities got it right when it comes to laboratory experiments or research projects during the degree: they make students work in pairs.

I studied physics and we worked in pairs. The real reason for that is that when you have X students and X/2 experimental setups, you have to pair people together.

I paired with a friend and it was fantastic: one week he would do the whole work, and the other week it would be me. (The TA was fine with that.)

Why? Because we hated to actually work in pairs. It was much more productive to do the whole work yourself because it was easy for us and we did not need to discuss the outcome (the experimental part was extraordinarily easy compared to the theoretical one where we often discussed the topics to try to understand).

I simply cannot imagine doing my PhD with someone else on the same topic. We would literally stomp on each other feet, or be constantly diverged from our thoughts by the other person. Discussing the progress is fantastic, once every few weeks or so.

Please note that this has little to do with being good or not. It is rather a personality trait.

I will also add that one dinner with a friend during my PhD completely changed its track because she suggested me something I did not think of, and it was revolutionary (for my thesis and a tiny little bit for the niche field I was working in). I praised her in the thesis even before my supervisor.

---

Fast forward to yesterday, 20-30 years later. I am now in industry and was working with one of my directors to remodel his team. One of the questions I individually asked the three best people in that team was "do you want to work together (as in "at the same time") on a problem, or do you prefer to work on it by yourself?"

The answer from each of them was a clear and sound "by myself". I asked why and they said that they genuinely like the other two (which is true), that they know that they are excellent (they are) but that they would be more productive concentrating on the problem by themselves and discussing/being challenged afterward.

Answer 7

In addition to what others have said about collaboration within the broader research group being more important:

• Ad-hoc pairing on a sub-project can be really useful and get results. I've got several co-authorships that way, including my most highly cited paper, and another that actually involved 3 postgrads from 2 groups working rather closely together.
• Forming these ad-hoc pairings (whether with other postgrads or postdocs) is a skill to learn in itself.
• Two postgrads, even if they can be recruited at the same time, are rarely working on such similar material that they can both work as a pair and get sufficient results for two PhD theses.
• Funding: even with a plentiful supply of willing students, getting PhD positions funded is hard to very hard for the PI/head of group.

This is largely from a physical sciences point of view, but informed by conversations I've had with other early career researchers across fields

Answer 8

I don't think that you get what a PhD really is - or at least what a PhD should be.

It is the consummation of independent, original and substantive research work that significantly adds to existing knowledge in that field and is deemed worthy of publication. It is also a qualification (though hopefully not the only one acquired) to teach in the discipline involved at university level.

Were universities to allow pairings to do a PhD programme - albeit where each individual looked at somewhat different aspects of a research topic - then how could examiners (or still less employers) fairly choose between those with paired doctorates and those with individual ones ? It's an everyday observation that operational compatibility between partnered workers is worth far more than the sum of their individual capabilities. A luckily paired individual could vastly overachieve and end up appointed to a position far beyond his/her individual capabilities.

Universities have always been individualistic research environments. This may not suit some of us who may like a real sense of combined effort in our workplace rather than the more abstract (and yes, at times, lonely) idea of contributing an individual effort to some indeterminate final whole. It may not suit efficiency maniacs in the Department of Education. But that is the nature of learning - we can really only see things coherently when looking at them our own way or from what has evolved from looking at them our own way. Cogging insights from other perspectives and jamming them in amongst our own will never lead to the clear coherent understanding needed to teach a class. It's not for nothing that a doctoral programme graduate is deemed a Doctor of Philosophy, rather than Doctor of Arts, Doctor of Science, etc. A programme forcing us to dig deep into our own thinking on any topic, questioning it quite often, squaring it with our experiments and those of others, arguing our hypotheses at seminars and walking the lonely walk between work and home every day - all this cannot but deepen our personal philosophy.

As things stand one could sometimes say that 'pairings' of a sort exist already unofficially: many PhD students just do the experiments suggested by their supervisor; many pick the brains of other students/staff and shamelessly use these ideas as their own; and many are inclined to misuse others to prioritise their claim on limited resources/equipment time - or even in getting personal attention. But this is down to lax attention to people management and poor personal example by lead academics.

The socio-professional benefits of pairing you mention - PhD candidates occasionally using each other as sounding boards, helping younger research students with techniques, analysis, etc - these are all things that any decent research group should foster. Departments where academics themselves cooperate tend to produce research groups like this; those where they do not will not have it. I think the happy medium is where the latter cooperation is freely given but every PhD student accepts his/her own obligation to do their own individual investigations. Yet please bear in mind that no one in any job can expect personal support from those around them at work: that's our own business and our own responsibility to find outside of work.

Your assertion of faster progress with research by pairings seems to me more like faster development/rationalisation of existing ideas rather than faster progress towards new fundamental insights. If a pairing does produce a fundamentally new insight, I'd be inclined to see it as the work of one mind in the pairing aided by the human support of the other one.

Answer 9

While it might be viable as an option, if you can figure out how to deal with problems such as one member doing the lion's share of the work, it wouldn't work for everyone.

First, and fairly obviously, is your university, department, and research lab large enough that there are likely to be two people at the same level who want to work on the same idea?

Second, perhaps not so obviously, is that not all people are alike. Even the statistics you quote show that: there are plenty of people who DON'T experience those symptoms of loneliness, anxiety, & depression, or at least aren't bothered by them. By the same token, there are people who very much dislike working closely with others. Force those people to do so, and odds are they'll soon be the ones reporting psychological problems.

Answer 10

There is another aspect to this: Lack of a functioning research group...

As a result you find a topic "dumped" on you as some funding was available and you get to pick up the pieces trying to figure out what to do... Then once you finish your PhD, you leave and with you all the knowledge you had. The person who comes along a few years later once again has to start from scratch...

I think pairing up people can also be difficult with respect to making sure that your PhD is your work as well as with respect to "ownership" on publications, even more so given how some people seem to be most intent to just profit for themselves rather than truly share/collaborate... However this can easily be resolved by having multiple people work on different aspects of the same niche as then these people can bounce ideas off each other while not being in direct competition and in fact benefiting even more as they bring different backgrounds and thus ideas to the topic.

(This is possibly more a comment than a real answer, but it is another, let me say experience...)

Answer 11

I think there is no proper reason and the scientific community loses many opportunities by sticking to this concept of letting PhD students mainly do their research on their own.

Of course, there is no doubt, that working in a team can be hard and challenging. And sometimes you have to make compromises and even admit that you were wrong and your collaborator had just better ideas.

But for the personal development, this - in my opinion - can only be beneficial as the PhD students learn how to collaborate, how to support each other and how you can benefit from specific skills of each of the team members.

---

Some of the previous answers argue that it will be difficult to decide in the end who is more responsible for the results, or who really made the work.

I think, as long as you do a good job, the people in your lab will notice, and they will see that you are dedicated to do your job, and that you are interested in doing research.

Moreover, the underlying assumption at this point is that science is rather competitive than collaborative.

As there is always a tension between competition and collaboration, the former one usually goes with separation as the latter one will lead to community. And as long-time separation is a breeding ground for anxiety and depression, having a strong community leads to security and a general well-being. (Admittedly, this is an assumption, but in my opinion a well-thought one.)

---

I, personally, was during my post-graduate studies in a working group with ~10 team members, where almost none of the PhD students could really talk with any of their colleagues as the topics of each of them were simply too far apart. Needless to say, that the atmosphere was kind of depressive. And most of them actually did not advise me to do a PhD for exactly that reason.

In essence, I think we only grew faster, personally and mentally, when we collaborate and encourage PhD students to do so.

Answer 12

I'll summarize my answer with: A PhD is not the moment for that.

Groupwork in undergraduate is extremely important because for those who do not follow academic careers it might be much more important to have social rather than technical skills. This is often said about many professions that are not highly technical, and at high-level management positions, technical knowledge is often of lesser importance compared to other skills, with great attention to managing and communicating with people.

During a Master's degree I'd expect someone to interact a lot with his/her advisor, co-advisor, and lab colleagues, maybe naturally find someone akin to what you describe as a research partner. Because at that moment the goal is to learn well all the stuff required to reach the edge of knowledge, to understand the state of the art, to learn how to stand on the shoulders of giants. So getting help and giving help within this context is good for everyone, as long as your master thesis remains, well yours. Some of this should still take place during a PhD, but some level of independence should also surface at this point.

I understand that a PhD is also a gateway that separates people who find research interesting in principle but are simply not fit for it (I find bodybuilding very interesting but have absolutely no place being a professional bodybuilder). It is a one-shot opportunity of demonstrating your capabilities to master the state of the art in a scientific subject, develop a rather complex project, execute it, make a verified contribution to the state of the Art, and last but not least take credit for it. You may disagree with this, but a lot of regulations in my country seem centered around those principles, such as you cannot ask for funding in a public agency unless you have a PhD (because the government does not trust you to conduct research with their money unless you have a PhD).

Once you've done it, society recognizes that you've shown yourself capable of:

• Understanding the state of the art in a scientific field
• Reviewing the scientific literature of your field
• Advising other students
• Managing research
• Reviewing proposed research articles

Now, you are eligible for several jobs in academia and in the industry where those skills are required, and guess what? You are in general not allowed to have a partner in all of those activities. The same way that the CEO position is very rarely shared between two or more people, any pure research or R&D project with grants, liabilities, and staff to manage require that one person to be ultimately responsible for some decision making, even if there is a group of people one can ask for guidance.

If your university forced or strongly incentivized students to pair up for PhD programs, can you really be sure that each member of every pair has demonstrated all such skills? I mean, if you feel at lost when you're not sure what to do with your research, you'll feel the same when you are no longer able to pay for a researcher's stipend because your project lost its grant, or if you have to fire an employee because the R&D project got canceled or a budget cut. The feeling is natural and not a problem, but handling the situation is the required skill.

To be clear, I'm not saying a PhD degree is the ultimate absolute proof of any of that, but it should be distinguished as a very strong indication of such.

Thinking back, I've seen many group projects in grad school where out of five students, one would do absolutely nothing, I've seen students who would often get help from one of the class top students and literally claim that they owed their (undergrad) degrees to that student. I've also worked with a guy in a company who was basically unable to do anything meaningful on his own.That guy enrolled in a Master's program and failed to deliver a thesis. Which was no surprise, as he always needed someone else to bridge gaps in his knowledge, plan things for him or review his ideas (preferably also correcting mistakes) and give him suggestions. When his advisor refused to hold his hand and act as a partner such as what you described, he just could not deliver. Do these descriptions paint him as someone who is really knowledgeable in his field, and thus able to manage research, advise students and review other people's work? Maybe you'd like to have a guy like this as a partner, but surely I'd prefer not to have him as an advisor or as a research manager. Let alone as a researcher in your university.

All that being said, once the PhD is finished, and this minimal demonstration of individual capacity is performed I'd see no problem in pairing up with another researcher for future work, no matter if for post-grads or tenured professors. Whatever is better than the sum of parts seems welcome if you ask me. However, once again, if either member of the duo decides to quit for a better position, gets hit by a bus or if the relationship sours, the remaining person is still expected in real life to be able to perform as an individual.

Answer 13

There seem to have been many good answers already and (sorry!) I haven't gone through all of them.

In my mind, you hit the nail on the head right at the beginning when you said, "they make students work in pairs. I loved my final year research project". In undergraduate work, you work in pairs, but usually in your third or final year... Paired work usually comes at the end of your undergraduate studies.

PhD is at the beginning of independent research. You're looking at it as if it comes after your final year of undergraduate studies. So if you did paired work then, then surely the next step should be more paired work or even group work. But PhD is (IMHO) the first step of another journey and not the continuation of your previous journey.

After you obtain your PhD, you'll have ample opportunities as a post-doctoral fellow and beyond to work in pairs or groups. Quite literally, if you choose to remain in academia then for the rest of your life, you'll work with others so much you'll get sick of it... The PhD is the beginning of this journey and it's this brief part when yes, you work alone.

(Of course, not every PhD program is like this. It varies from university-to-university and country-to-country. From personal experience, once you open up "group work" right at the beginning of a PhD, many people will abuse it. Sure -- perhaps you're a good person and you wouldn't dream of it...but many, sadly, would. It would be like allowing paired work and open book exams in first year of your undergraduate studies.)

Lastly, we should separate the PhD degree from the ability to do research. Anyone can do research without a PhD degree. So, two people (i.e., A and B) who have been brought into a lab to work together can be hired as research assistants (well, the term varies from country to country...but basically, the only qualification needed is an undergraduate degree) and produce a unit of research as a research publication with join co-authorship. So, no university (AFAIK) is stopping joint, collaborative work...