Five Strategies for Leading Interdisciplinary Research Teams

This post builds on the Integrative Capacity framework I developed with colleagues to examine what distinguishes scientific teams that produce genuinely integrated knowledge from those that produce parallel work that never quite connects. The strategies here are practical applications of that research, written for research leaders managing the day-to-day realities of multi-investigator collaboration.

The project kickoff meeting went well. Everyone was aligned on the science, enthusiastic about the collaboration, and clear on their piece of the work.

Three months later, a co-investigator discovers that another site has been conducting a parallel analysis using a dataset that another lab had already wrapped up. Nobody flagged it because no one realized the duplication of effort had occurred. The cost was four weeks, just because there wasn’t a conversation about the workflow and who was going to do what.

This is not just a communication failure. It is a coordination failure. And it is one of the most common ways that well-funded, well-staffed research teams lose time, trust, and momentum.

Leading an interdisciplinary research team requires a different skill set than leading a lab. In a single-investigator lab, you set the direction, you know everyone's work, and your authority is clear. In an interdisciplinary team, authority is distributed. Expertise is deliberately varied. The people you depend on most have other labs, other obligations, and other legitimate priorities.

The strategies that work in one setting can quietly fail in the other. Here are five that actually travel.

1. Design the handoffs before you need them

Most collaboration problems don't start with conflict. They start with ambiguity about who owns a transition.

Data moves from one site to another. A protocol gets modified. A sub-study yields a finding that changes the primary aim. In each case, someone needs to make a call, and in interdisciplinary teams, it is rarely obvious who that person is. In research I conducted with colleagues on leaders' integrative capabilities, one distinguishing feature of leaders who sustain effective, discipline-diverse teams is this: they build a coordination structure before the team encounters the situation that requires it, rather than reconstructing the process after something goes wrong.

Before your next project launch, ask each site lead to name the three moments in the work where they will most depend on another team. Then ask whether there is a named process for each of those moments. If the answer is no, you have found the places where the collaboration will likely break down.

2. Treat the collaboration plan as a living document

Most collaboration plans are written once, submitted with the grant, and never opened again.

This is a problem not because the document is unimportant, but because the work it describes will change. Co-investigators leave. Aims shift during the no-cost extension. A site that was secondary becomes load-bearing during certain critical stages of work. The team that submitted the proposal is rarely the team that finishes the project.

A collaboration plan reviewed annually does something a static document cannot: it forces the team to look directly at how the collaboration is actually functioning, not how it was intended to function. Where has decision-making drifted from what was originally agreed? Where are people absorbing work that isn't formally recognized? What norms formed organically in year one that nobody would have written down on purpose?

Reviewing the collaboration plan is not paperwork. It is a diagnostic.

3. Build conflict infrastructure before you have conflict

Conflict in interdisciplinary research teams is not a sign that something went wrong. It is a sign that people with genuinely different expertise and genuinely different priorities are trying to produce something together.

What goes wrong is the absence of a shared protocol for surfacing and resolving disagreement. Without one, small scientific disputes tend to go one of two ways. They get suppressed until they surface as something larger, or they get escalated to the PI as interpersonal problems when they were originally scientific ones. Either way, the team is now managing a relationship problem that started as a methodological disagreement.

Name two things explicitly with your team before the first point of friction: how scientific disagreements will be raised, and how final decisions get made when consensus isn't possible. Teams that have this conversation in the first month rarely need the harder version of it in month eighteen.

4. Make integration work visible

In most multi-investigator projects, someone is doing the coordination work that holds the collaboration together. Synthesizing inputs from multiple sites. Catching version-control problems before they become data problems. Translating between disciplinary vocabularies so a statistician and a clinician can actually understand each other's concerns. This work is load-bearing. In most evaluation systems, it goes unrecognized.

In work with colleagues at UCI's Team Science Acceleration Lab, we identified credit-sharing as one of the primary institutional leverage points for sustaining collaborative research over time. When integration work isn't named in budgets, authorship conversations, or promotion reviews, the people doing it burn out or move on. When they leave, projects stall in ways that take months to diagnose.

As a team leader, your job is to name this work explicitly, in meetings and in writing. Who is doing the integration? Is that role formally recognized? Are those people being trained for the work they are actually doing, or improvising infrastructure under pressure? The teams that sustain high performance over multi-year projects are almost always the ones where that question has a clear answer.

5. Protect the shared mental model

Interdisciplinary teams are assembled because no single discipline can address the question alone. This is the model's strength. It is also its primary maintenance challenge.

Over time, sub-teams develop their own language, timelines, and working theories about what the project is actually for. This is normal. It becomes a problem when those mental models diverge far enough that sub-teams are no longer working toward the same scientific goal. Research I conducted with Theresa Lant on integrative communication in interdisciplinary teams identified this as one of the central functions of leadership: not just managing each sub-team's progress, but actively maintaining a shared orientation toward the collective question. When that orientation drifts, the sign is usually subtle. A co-investigator frames a finding as if it were a different primary aim. A trainee from one site presents work that doesn't connect to the project's overarching question in any way they can articulate.

The remedy is not more meetings. It is a periodic, structured conversation about what the team is trying to produce collectively. Ask each sub-team, once a year, to explain how their current work connects to the project's core question. The answers will tell you quickly whether the shared mental model is holding.

The common thread

None of these strategies requires a personality change or a team retreat. They require infrastructure: clear handoffs, a living collaboration plan, named conflict protocols, visible integration roles, and a protected shared understanding of what the work is for.

Research teams that build this infrastructure before they need it don't avoid problems. They avoid the specific class of problems that waste the most time and do the most damage: the ones nobody saw coming because nobody designed the system to catch them.

If you want a structured way to assess where your team's integration is holding and where it needs attention, the Research Team Integration Diagnostic is a good place to start. If the gaps are significant enough to warrant focused development work, the Scientific Lab Accelerator is designed for that.