Bigger isn't Better, but More is Different

Research labs in the UK are beginning to open back up, and many groups are starting to figure out how to combine social distancing with safe and productive lab work. This creates challenges for all labs: how can there be enough people in the lab to be safe, but not so many that social distancing is impossible? How can projects, reactions, analysis, prep-work be prioritized fairly for everyone? Is it fair to schedule lab time outside traditional working hours? Some of these challenges might be particularly acute for larger groups. This has to the suggestion that members of larger groups suffer inequities not just as a result of COVID-19 but as a general matter and further to the idea that funding bodies could cap group sizes to prevent such inequities from occurring. As a post-doc working in one of the largest groups in the country, and as someone who graduated from a relatively small group in the US, I cannot understand how limiting group sizes would improve career outcomes for anyone. Not because bigger groups are better, but because bigger groups are different.

The lab I work is has more than 60 people, with 26 post docs, 17 graduate students and 6 technical staff. It’s giant. Due to its size, it operates completely differently from the lab that I graduated from, which had 1 post doc, and 2 PhD students (myself included) when I started my PhD. The lab I work in now is organized into 6 research teams that pursue specific research agendas and are led by 2 post-docs with domain expertise who serve as team leaders. As a team leader myself I help organize and facilitate research for 11 people, posts docs and grad students alike. The experience I’ve gained from over-seeing so many different research projects is invaluable. How many other post docs can go into faculty interviews and demonstrate they know what it takes to manage an entire research team? The structure of this large group has been instrumental in helping me progress in my own career. If I get the chance to become a PI myself, I will be better at training and supporting students because of opportunities that could only exist in a larger team.

Having two team leaders creates a more diffuse and transparent power dynamic. In all my interactions with the team my direct peer can ask me to explain myself, and if someone on my team is upset with my behavior there is someone they can go to. On top of the team structure, which organizes research agendas, there are four sub-groups within the lab that help keep track of individual career trajectories through the group. Sub-group leaders are another mechanism of transparency and accountability, they help ensure sure individuals get what they need out of their post-doc or graduate experience. Finally, the lab is lucky to have an involved group coordinator, who helps organize projects between teams, and works with the team leaders. If any member of the lab has an issue, either a technical research question, a career concern, or an inter-personal issue (such as bullying), there are 3-4 people in the lab they can talk to, who are not their PI. Those 3-4 people are empowered to raise the issue to the PI, or discuss with other leaders in the lab, while maintaining the anonymity of the people involved. Its not a perfect system, but it creates a lot of options, many of which would not exist in smaller labs.

Are there limitations associated with bigger labs? Absolutely. One thing you cannot do in a large team, is work on completely independent projects. It just doesn’t make sense to pursue your own research agenda when there’s so much team infrastructure. That doesn’t mean members can own ‘their’ project, just that each project is part of a larger vision. Within the scope of that larger vision, every individual project might seem incremental but in aggregate they are more exciting. For example, when I started in the lab, I was fresh out of grad school and I had ideas I wanted to run down myself. My PI had other plans in mind. Initially I was extremely frustrated that he kept pushing me to model a system that seemed simple to me. After some strong encouragement I got to work on the project he’d laid out for me and the pieces of the bigger vision started to come together. I started working with a senior member of the group who had characterized the system experimentally, with the help of other people who had left the group. The data they had collected was really rich and the core conceptual idea we narrowed in on was really exciting. Project was more interesting than I had originally thought, in large part because there was more to it than I could understand at the beginning. The larger group structure had created an opportunity for me to participate in a project that had been on-going for many years. In doing so, I got the chance not only to contribute to an exciting paper, but applied skills from my PhD to an entirely new disciplinary area. Sure, it wasn’t my idea, but working on it helped me grow as an interdisciplinary scientist in meaningful ways.

When I was starting as a PhD student, I enjoyed long (>1 hr) and regular (weekly) meetings with my PI for the first few years of grad school. If that’s the kind of environment you’re looking for, a group with 60 people, is not for you. But that doesn’t mean you can’t find individualized attention. As a team leader I can easily meet with team members for extended technical discussions on a regular basis, which can help compensate for a PI who can only check in for a few minutes most weeks. In addition, working in larger teams means that more of your peers have given you feedback on ideas, experiments, models and analyses by the time your PI looks at your work, which can help build confidence in your results. In fact, research has shown that engagement from post-docs and senior PhD students better predicts skill progression than PI engagement.1

One thing is clear, its not obvious that larger groups should be seen as a priori more productive. Productivity in sciences can only be crudely quantified using controversial measures like number of publications, or number of citations. But from what we can measure, it’s clear that big teams and smaller teams are different. For example, research shows that bigger teams publish more work that can be characterized as incremental, while smaller teams publish more work that appears innovative.2 These differences are a good sign! Philosophy of science tells us that progress toward a deeper understanding of reality requires both incremental progress and innovative, disruptive new ideas. Any attempt to stymie one of these modes of progress would be counter productive to the scientific endeavor.

Finally, in the face of the COVID-19 crisis, it’s not obvious that members of larger teams will suffer disproportionately to smaller teams. Yes, limited lab space presents a challenge, but team projects mean that most individuals don’t need to be at the bench to move everyone’s projects forward. Having a diverse set of experts working towards the same goal means that everyone can contribute where their time is best spent, maximizing limited time while social distancing measures remain in place. Bigger teams might face more problems during the crisis, but collectively they can generate more solutions. They won’t have an easier time through the crisis, but a different journey.

Make no mistake, contemporary science faces serious structural and systemic problems. The inequalities and injustices that exist in society also permeate the institutions of science. Reducing the effects of these injustices will require novel new paradigms in the way science is funded, communicated, and conducted. Those new paradigms won’t come from limiting the ways groups can be organized, but instead encouraging a diverse ecology of research teams.3

  1. Feldon, D. F., Litson, K., Jeong, S., Blaney, J. M., Kang, J., Miller, C., … & Roksa, J. (2019). Postdocs’ lab engagement predicts trajectories of PhD students’ skill development. Proceedings of the National Academy of Sciences, 116(42), 20910-20916.
  2. Wu, L., Wang, D., & Evans, J. A. (2019). Large teams develop and small teams disrupt science and technology. Nature, 566(7744), 378-382.
  3. Clauset, A., Larremore, D. B., & Sinatra, R. (2017). Data-driven predictions in the science of science. Science, 355(6324), 477-480.