Research shows that studying in groups help students learn more effectively.

If this is the case, why don't academics also conduct research in groups? There are already research groups, but in all the cases I've seen, each individual member of the research group works on a separate research question. If the current question is "reproduce the results of this paper", usually one individual member works through the paper alone (i.e. conducts research alone) and reports the results to the others.

It seems likely to me that the benefits of studying in groups should also happen when conducting research in groups. With someone else that's intimately familiar with what you're doing, you can understand papers better, catch coding bugs quicker, and so on. The sum is greater than the parts, and one can achieve more in less time. Nonetheless, I don't think I've ever seen a single PhD project with two assigned PhD students.

What is the rationale for not conducting research in groups?

Edit: I'm referring to the kind of group work that undergraduates might do when working together: one is next to each other when working, and frequently bounces questions off each other. Computer programs are written together such that both partners know exactly what each line is doing, experiments are conducted with two pairs of hands (or more), and so on.

My experience is that at research level, this doesn't happen. Computer programs can be written by many people, but usually each person is responsible for an individual section. Others know what the program is doing on a high level, but are not familiar with the nitty-gritty of other parts of the code such as what each line is doing. Expressions of the "I see what the code is doing, but I would not have written it like that" are common. Similarly, the dirty work of experiments (e.g. aligning mirrors in an optics experiment) is carried out by one person. Others might know what the experiment is trying to do, but don't get personally involved unless the main experimenter gets stuck (or if there is some kind of spectacular discovery).

  • 32
    But they do. Most paper are co-authored in my field, and enourmously large collaborations exist notably in physics. In maths, look up "Erdős-number". And then, of course, workshops etc. are a thing in most fields. – henning May 6 at 6:52
  • 2
    We do research in groups, especially in experimental fields. About the statement "these authors contributed equally", for instance, in my field author contribution statements are quite unusual. – Massimo Ortolano May 6 at 7:22
  • 5
    Students doing group work ie projects for assessment often allow the "good" ones do all the work. Seen this with some students in a "good" group who then don't do so well in the final exam. Also, is this for all subjects? – Solar Mike May 6 at 7:28
  • 2
    @MassimoOrtolano how much of a group effort is it? For example, I've coauthored a paper with someone where both our programs produced the same results; however I have never examined his program and he has never examined mine. We coauthored the same paper, but we effectively worked alone. – Allure May 6 at 7:30
  • 2
    @Allure Not even close. Large software projects are never that nicely partitioned, even if you're lucky enough to have exactly the same set of programmers from start to finish. – JeffE May 6 at 12:54

As the comments indicate it is quite common for researchers to cooperate, but the extent to which that happens differs a lot between sub-disciplines.

However, I would argue that your analogy is false. Studying is all about acquiring existing knowledge or skills, while research is all about creating new information. What works well for one type of task does not necessary also work well for another type of task. So maybe there is value in cooperation in research or maybe not, but the study you quote does not help us answer it.

Reaction to Edit of question

When I collaborate I like to find someone that complements my abilities. This way collaboration in research tends to favor specialization. There can be some value to let two persons do the same task in terms of quality control, but that is also very expensive. In my case the expected benefit has always been nowhere near the cost, so I have never done that.


I've no idea what field you're in but most research in most areas of STEM subjects is done by groups of researchers so, in the generality that it's stated, the question is based on a completely false premise.


The benefit of teamwork derives from division of labor, knowledge, and viewpoint. Unlike a team of draft horses, an academic or work team should not be evenly matched. Adam Smith told us 300 years ago that specialization would make the process of pin-making fifty times more efficient. A team should divide the work into daytime and nighttime data collection, writing the database code, writing the front-end code, and a specialist to write the grant, create the figures, and edit the report. They should gather together to help each other over problem hurdles and suggest better approaches.

The most innovative of teams may well be pairs: Hewlett and Packard, Page and Brin, and Jobs and Wozniak. But a team of two can't do everything in a finite period of time. So you need to add people. Those people are each experts in their own field; that gives economies of scale. Empirical evidence shows that economies stop rising after team size reaches 250. Best to put the next set of people under a new roof. It's tough to recognize more than 250 faces. Getting cc'ed by more than 250 takes more time than it's worth.

As to the practice of dividing a class into teams of four, I suspect that it reduces the effort a prof needs to grade the projects. Speaking as a hiring manager and behavioral interviewer, I found questions about group work rich ground. I was instructed to listen for certain evidence in the response to those questions. Lots of time I heard that Billy Bob didn't participate in the group but to eat the pizza and get the grade. Fine. But if from the interviewee grousing was all I heard, or that she took up the slack, a black mark went on the paper. Taking up the slack is not what team work is all about.

As to studying in groups, I think it fills the same purpose as recitations -- understanding that your classmates are encountering the same stumbling block you are. The group allows communication in terms common to the group, even if the prof doesn't use the same analogies and examples. If the group dynamic is such that subject material is divvied up, with an expert on each, that's brilliant. There is no better way to learn something than to have to explain it to another.

  • 1
    "A team should divide the into daytime and nighttime data collection" Remember that not everyone is an experimental scientist. And not every experimental scientist has long-running experiments that need 24/7 data collection. Indeed, such people are probably quite a small minority within academia as a whole. – David Richerby May 6 at 20:34
  • 2
    Yeah, this is quite field specific. daytime and nighttime data collection, writing the database code, writing the front-end code, and a specialist to write the grant, create the figures, and edit the report: Here in pure math, none of my research projects have involved data, collection, databases, code, or grants. I think one of my papers had a figure, but it was a while ago. And we don't call our papers "reports". – Nate Eldredge May 6 at 20:49

Please check the author list of this paper, which runs over nearly 3 journal pages towards the end of the paper, as an illustrative and representative example that people do work in groups. Certainly not everyone duplicates the work of others - this is inefficient and would require unimaginable training to get the proper expertise to get everyone to be proficient on every detail of such an experiment - but the LIGO collaboration can achieve outstanding results only by leveraging the expertise of hundreds of different persons.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.