Empires, credibility, and a happy workplace atmosphere

By Joern Fischer

Bigger is better in academic reward structures: more grants, more students, more papers, more impact factor. But is there a limit to how big research groups (or “labs”) ought to be? What are the pros and cons of big, medium, or large research groups in terms of producing quality science, scientific credibility, and a happy working environment?

I figured thinking through this could be interesting, but as you will see, I didn’t get very far! Let’s start with very small research groups. Small groups can produce very good science if the individual researchers are very good. This might be especially the case for subjects where it is not necessary to draw on many different kinds of expertise, or where individuals work largely on their own. Here, it’s really largely the quality of the individual researchers that matters. If they’re very good, the science produced will be very good – and nobody would have any doubts about its credibility. On the other hand, if the individuals are weak in some areas, there are few opportunities to buffer one another in such a situation (e.g. if there are three people, and none of the three is good at statistics, then the statistics ends up … well, just not very good!). And complex subject matters, which require multiple different perspectives probably can’t be dealt with very well in extremely small research groups.

Medium-sized groups then … the most obvious advantage here is individuals can buffer one another more effectively, and there are more opportunities for mutual learning and collaboration. There is also the chance of having the beginnings of a “critical mass” of people who collectively can push forward a common approach or idea. In medium-sized groups, communication within the group is still quite straightforward, and the potential for people to be happy (in a social sense) is quite high.

What about really big groups? Well, they are clearly the most prolific. A glance at google scholar suggests that several of the world’s leading conservation scientists now produce over 50 publications a year, for example. So, is this the way to go, something to be frowned upon – or just something that needs to be managed very carefully? To start with, I guess the benefit of having a critical mass cannot be denied in such big groups. They are basically “centres”, though often somewhat more coherent in subject area because they are run by a single senior academic. On the downside, various risks also increase in really large groups. There’s a risk that the quality of the individual researchers in the group can’t be consistently high – it’s more difficult to hire large numbers of truly excellent people all at the same time than to hire a small number of truly excellent people. And then there’s the risk that people will be impressed by, but at the same time cynical, about very large groups: can anyone really contribute meaningfully to more than 50 papers a year? (I’m not sure, but I am sure that many people would say “no”!) And finally, there’s also the risk that overall group cohesion is lower, partly because communication within a large group is much more difficult. And so, very large groups in reality often split into a number of smaller sub-groups.

In the end then, if such sub-groups are working well, there is no reason to believe that large groups should be inherently less pleasant to be part of than small ones. They might even be particularly nice because they have a critical mass of like-minded people. If their governance is organized sensibly, perhaps large groups are in fact the best research environments … ? Or perhaps it’s small groups for some purposes, and larger ones for other purposes?

As I’m approaching the end of this blog post, I have not reached a definitive conclusion on what to make of academic group sizes. Perhaps size is just not an interesting feature in its own right: perhaps it’s how the group is run that matters in terms of the quality of the science and the workplace atmosphere … I’d be interested in people’s thoughts! What do you think, and what are your experiences with small versus large research groups?

14 thoughts on “Empires, credibility, and a happy workplace atmosphere

  1. It would perhaps be helpful (in terms of a discussion) to outline in more concrete terms what small, medium or big means (I suspect this is a very personal thing).

    Also it is probably necessary to know if a big group is actually 3 medium sized groups (each with their own leaders) with a titular head*

    *Titular is not really the right term here, but it is close to what I mean, sounds funny and it is very close to home time, so it will do.

    • Hmm, I guess small means like no more than 5 people, medium might be up to 10, and over 10 might be large … with 20 being definitely very large? I’m counting “core researchers”, i.e. typically PhD students and postdocs.

  2. Group size and group dynamics has had a lot of research attention – I just quickly found this nice little overview from the Wharton Business School: http://knowledge.wharton.upenn.edu/article/is-your-team-too-big-too-small-whats-the-right-number-2/ It considers group size, sub-groups, the benefits of diversity, fitting team size to task, etc.

    From personal experience, I would say that it depends on who is in the group and who is in charge, and on the broader environment that group is situated within (access to input from other groups).

    And, it depends on what we counting as effective (as you refer to in your title). I suspect that the optimal group size would be different between two groups where one is effective if it produces many papers, and the other is effective if it produces sound research and satisfied researchers.

    • Thanks Julia, that’s a really interesting link! It suggests one ought to consider the task at hand; not under-estimate face-to-face communication; and if groups are large, perhaps plan for sub-groups. And some degree of common understanding within sub-groups (as opposed to super-high diversity) is also useful, according to that. All makes sense to me … I found it a nice summary of some key things to think about.

  3. I really like this topic. Thus, I try to share some of my thoughts. And I would like to approach it from the perspective of leadership. A leader (which could refer in this example to the lab head) has a vision: Let’s say to move forward urban ecology. What I learned about leadership (having said that, I am still a beginner) is that as a leader you know exactly your strengths and weaknesses. For instance, you are really good in seeing the bigger picture. However, you might lack e.g. in GIS or statistical skills. Thus, you seek exactly those people with skills that would fill these gaps. Once you have filled all the gaps that you consider to be relevant to move forward with your vision it would be, after my point of view, the right group size. However, to be productive, you would need to motivate your group of course as well. E.g. organize writing retreats, encourage to share thoughts among each other, collaborations, further training, etc.. Maybe this is too simple. But I guess this is how I would try to approach to produce quality science, scientific credibility, and a happy working environment.

  4. I have worked in and around small and large research groups in my career and the ‘happiness’ of the group largely comes down to the type of leader who stewards the group. Command-and-control leaders of big groups foster competitiveness and high outputs (of average quality) and group dynamics foster considerable ‘unhappiness’. Leaders who are happy to share some of the power and have sub-groups that run somewhat autonomously seem to do better. So, the size of the group is one thing, the style of leadership and the group structure are other important ingredients. All of which is really just my opinion.

    However, the quote below from a very recent Biological Conservation editorial puts some numbers under the spotlight (see their article for a breakdown of those numbers). They support Joern’s opening statement: “Bigger is better in academic reward structures.” The rich get richer, the poor find it hard to compete.

    “Another noteworthy feature of the data set is that 4 of the 5 authors with the most publications in Biological Conservation live in Australia. Seven of the 10 top most-cited authors of papers published in Biological Conservation live in Australia. These prolific authors are the leaders of large research groups characterized by many graduate students and post-docs. These groups are productive because they submit many papers each year, and not because their papers have a higher rate of acceptance.”
    Richard Primack and Lucy Zipf (Editors in Chief, Biological Conservation)

    • Hi David — interesting. This brings in the perspective of who you ask… So, a big team is effective for the person running it! But not necessarily more effective for the people who are part of it. My interim conclusion on all this is that if one were to run a big team, one ought to actively manage for sub-teams, and see those as the primary unit of being “productive”. Cheers — J.

  5. Great post, and it is on a topic that has bothered me a lot as an early-career academic at a small university.

    Someone who I respect a lot (I won’t name-drop here…) once told me that the perfect size for a research group is seven people (from a mix of starting PhDs, senior PhDs and post-docs). According to him, seven is small enough to ensure regular face-to-face interaction, but still big enough to create intellectual critical mass. It made sense to me at the time.

    But now, one year into my first academic position, I realise that the size of an effective group depends on the quality of the researchers you are working with. You can allow your group to grow much faster if you are working with really smart, dedicated and autonomous people. Unfortunately, if you can’t attract such people (as is my case as an inexperienced scientist at a small uni), then it is better to rather keep the group small and dedicate more attention to fewer students.

    I don’t know if this is a good strategy – ask me again in five years time – but it is what I’m going for.

    • Hi Falko — for what it’s worth, I agree with you! I think you’re better off having few people who you really believe in (and like to work with) than “diluting” a group for the sake of size; a bigger group per se is not a meaningful goal for its own sake, I’d say. Cheers — J.

  6. Excellent and fascinating discussion here. For once I haven’t very much wordy to add!

    The brief note I would make is it seems like intermediate size could be an optimal balance for the interesting purpose of balancing expertise and research productivity with the justice- pedagogy- humanistic-focused outcome of being able to incorporate people who *are not yet excellent, but have substantial potential.* I suppose one could that this is resolved by the simple tautology of only incorporating people with such potential, but I think doing it right involves taking risks on people who will not “obviously” be able (but are willing!) to rise to the occasion. (The only way to break a cycles of lower opportunity would seem to be to go beyond offering opportunities to the most exceptional of the marginalized…)

    With a medium-sized group, there may be enough redundancy that good work continues even as some members are “taking” more time than they’re “giving” at present, but small enough size that such a person (or people–there can be importance for critical mass here too) are not lost in the shuffle and get sufficient mentorship to excel (socially and scientifically).

  7. The question has been quite extensively studied – see paper by myself and two undergrad students (Cook, Grange and Eyre-Walker (2015) PeerJ https://peerj.com/articles/989/) and references therein. Most analyses of data have found that the number of papers published by a group increases with group size, but the relationship is one of diminishing returns; the mean impact factor and number of citations also increase with group size but only very weakly.

    In our analysis we found that the number of papers over a 5 year period was linearly related to group size, but that the intercept (the productivity in a group composed only of a PI) was positive and much greater than the slope; hence the productivity per person drops as group size increases giving the impression of diminishing returns. I say it gives the impression of diminishing returns because the real story in the analysis is the high productivity of groups with only a PI – i.e. adding a post-doc to a small group seems to lead to a similar increase in productivity as adding a post-doc to a large group. We argue that from our results that we should increase the number of PIs and decrease the number of post-docs and PhD students because PIs by themselves or in small research teams, seem to be more productive than post-docs and PhDs.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s