Collaboration in research is rarely as clean as the org charts and grant personnel profiles suggest. As biostatisticians, we often sit at a unique intersection - technically indispensable, yet often operating at the edges of decisions that shape the work. When things get difficult, and they will, the way we navigate those moments can matter as much as the quality of our methods.
Most of the tension we feel in collaborative research isn’t about us. It’s about everything happening around us that we often can’t see or influence.
The Instinct to Interpret — and Why It Misleads Us
When something feels off in a working relationship, we often reach for explanations quickly; and we tend to reach for the ones that center on ourselves: They’re frustrated with me. They don’t trust my judgment. They think someone else could do this better.
Sometimes those explanations are worth examining. But I think more often, they aren’t accurate and acting on them quietly shapes how we show up. We become a little more careful, a little more distant, a little more focused on being covered than on being useful.
Sometimes we withdraw in meetings because we’ve convinced ourselves that a project leader—or even the team—doesn’t value our input or may be working against us. Left unexamined, these feelings can grow, leading us to act defensively, such as cc’ing supervisors on emails as a subtle form of self-protection. A more constructive approach is to address concerns directly through open conversation. In many cases, others are simply preoccupied with competing priorities that we may not see or fully understand.
The instinct to protect ourselves (often passively) is natural. But it can crowd out the more useful instinct, which is to stay curious about what’s actually going on.
Know your leverage and who the relevant decision makers are.
We have influence in research collaborations. The work doesn’t move forward without us, and our teams generally know that. This influence can feel empowering when things are going well and a sort of quiet leverage when they aren’t.
Leverage is the ability to escalate, to push back formally, to have a meaningful influence on the outcome of a decision. There are moments when using it is the right and necessary thing to do, but it’s worth sitting with it for a while before reaching for it. Often we are just not positioned to influence the outcome.
Consider a situation where we feel a collaborator isn’t respecting the time required for careful work where deadlines keep shifting, requests come in at the last minute. After months of this, the concern gets surfaced sideways, mentioned to a supervisor or in a meeting before ever being raised directly. The collaborator (or the biostatistician) hears about it secondhand. The relationship, which had been functional if imperfect, becomes guarded on both sides.
Now consider a different version. We ask for thirty minutes: “I want to make sure I’m supporting you as well as I can. Let’s talk through how the timeline is working from your end.” Maybe the PI shares that a an unexpected deadline was realized and they’ve been quietly pushing the work forward ever since. The conversation doesn’t solve everything, but it changes what the problem actually is. And it keeps the door open.
Leverage used too early forecloses conversations that haven’t happened yet and some of those conversations could change everything.
Looking for What’s Actually There
Hard situations have textures. When we slow down enough to look, we usually find something more complicated, more workable, than the story we started with.
When Personalities Clash
Some PIs are blunt in ways that sting. Others are so deferential that you never quite know where you stand. Some go silent under stress; others flood your inbox. Early in a collaboration, these differences can feel like friction. Over time, they tend to reveal themselves as information about how someone operates; not about how they feel about us.
Terseness in writing is not always coldness. Brevity in feedback is not always dismissal. Most communication styles that initially read as difficult become more navigable once we stop expecting the other person to communicate the way we do, and start paying attention to what they’re actually trying to convey.
When Workload Is the Real Story
A PI who seems disengaged, slow to respond, or inconsistent in their priorities is usually not indifferent; they are likely overwhelmed. PIs carry a weight that is largely invisible to most collaborators: grant cycles, teaching loads, committee obligations, mentoring, clinical responsibilities, departmental politics. The bandwidth left over for careful, unhurried collaboration is often genuinely thin.
When that’s the context, the more useful question becomes: What can we do to make this easier for them to engage with? Sometimes that means shortening an email to three sentences. Sometimes it means offering a concrete recommendation instead of a list of options. Small adjustments, but they can change the rhythm of a collaboration meaningfully.
When Funding Shapes Everything
Research happens inside grant structures, contract timelines, funder expectations, and institutional priorities — and most of those constraints are not visible to everyone on the team. When a PI makes decisions that seem inconsistent or hard to follow, there is often a funding conversation in the background that hasn’t been shared.
Asking about those constraints — not to challenge, but to understand — is one of the most useful things we can do. Is there context I’m missing about where this project needs to go? That question, asked genuinely, often surfaces the real situation. And once we understand the real situation, we can actually help.
When the Budget Runs Out and We’re the One Who Has to Move
A PI reaches out and asks to meet. The tone is a little more formal than usual. And then they tell us: there isn’t enough funding left in the grant to keep us on at our current effort level. They need to move us off, at least for now.
The first instinct is often to feel betrayed — I’ve been here from the beginning. I know this data. I helped shape this work. And underneath that, something harder: How did this happen? Didn’t they see this coming?
It’s worth understanding the structural reality here, because it shapes what we’re experiencing. PIs are responsible for grant budgets, but those budgets involve a lot of moving parts — indirect rate changes, no-cost extension timing, personnel transitions, co-investigator effort — and forecasting errors happen even with good intentions. Key personnel, unlike co-investigators, have limited formal visibility into those financial decisions and limited structural recourse when they change. That isn’t a failure of goodwill; it’s a feature of how grant-funded research is organized. Understanding it helps us respond more clearly.
The PI delivering this news is often carrying real discomfort about it. They may not have communicated early enough. They may be scrambling for solutions. Whatever led to the situation, the conversation in front of us is still worth having: Is there a path back onto the grant? Are there other projects where my effort could be supported? Can we document the work completed so it isn’t lost? These are fair, professional questions, and how a PI engages with them tells us something useful about the relationship going forward.
One thing worth paying attention to — not with suspicion, but with clarity — is the framing of the conversation. Budget-driven transitions are sometimes presented in performance-adjacent language: concerns about fit, or efficiency, or pace of work that hadn’t surfaced before. If that framing feels inconsistent with your actual record on the project, it’s worth grounding yourself in the documentation you’ve kept — decisions made, work completed, timelines agreed to — before internalizing feedback as a signal about your abilities. A paper trail isn’t paranoia. It’s context.
If this pattern repeats across projects — if budget decisions consistently happen to us rather than with us — it may be worth thinking structurally about how we engage in future collaborations. There is a meaningful difference between being listed as key personnel and being named as a co-investigator. As a co-I, we are named on the award, visible to the funder, and part of the formal record of the project. That visibility doesn’t eliminate all uncertainty, but it means that decisions about effort and budget are made with us rather than around us.
Pursuing co-investigator status is most effective when the conversation happens early — when a grant is being developed, not after it’s been awarded. If we are shaping the statistical design, contributing to specific aims, and carrying analytical responsibility across the life of a project, it is fair and appropriate to name that contribution clearly from the start. A PI who genuinely values the collaboration will usually engage with that conversation seriously.
What tends to be least productive in the immediate aftermath of difficult news is letting frustration close down the conversation. The PI sitting across from us may still be the person best positioned to help us find a stable next step. The goal isn’t to minimize what happened — it’s to stay clear-headed enough to protect our interests while keeping the door open.
When the Data or the Model Is the Hard Part
Sometimes the difficulty in a collaboration has nothing to do with people at all. The data doesn’t behave the way anyone planned. The model surfaces something inconvenient. A key assumption doesn’t hold. These moments are genuinely stressful, and that stress tends to find a target.
The most useful response in these situations is not defensiveness and not detachment — it’s steadiness. A PI who is worried about what the results mean for their study doesn’t need an explanation of why it isn’t our fault. They need a collaborator who can think clearly alongside them about what to do next.
Some Things Worth Practicing
These aren’t rules. They’re habits of mind that tend to be useful when situations get hard.
- Ask one more question before drawing a conclusion. “What’s driving the timeline on this?” or “Is there context I’m missing?” often opens up the conversation entirely.
- Name the constraint, not the person. “I want to make sure we’re protecting the integrity of this finding — can we talk through the tradeoffs?” lands differently than “That approach isn’t going to work.”
- Offer a path, not just a problem. When we raise a concern, coming with at least one idea for how to move forward signals investment in the outcome, not just the complaint.
- Find perspective, not an audience. When we’re struggling, talking to a trusted senior colleague — not to vent, but to think — makes a meaningful difference in how we show up afterward.
- Document decisions, not grievances. Keep clear notes on what was decided and why. It protects the science and keeps the focus where it belongs.
On Knowing When to Escalate — and Who Actually Matters
There are situations that require escalation — when scientific integrity is genuinely at risk, when ethical concerns arise, when a working relationship becomes harmful despite real effort to repair it. Recognizing those situations and acting on them clearly is part of professional integrity.
But escalation done well comes from clarity, not exhaustion. Before escalating, it’s worth asking: Have we said, directly and clearly, what the problem is? Have we given the relationship a genuine chance to work through it? Can we describe the concern specifically, without it becoming about the person? If the answer is yes, advocating firmly and through the right channels is not a failure of collaboration — it’s part of taking the work seriously.
And one more thing worth saying directly: be thoughtful about who you escalate to. Not everyone we might want to talk to is actually in a position to help. Bringing a concern to someone who has no decision-making authority and no meaningful influence over the situation doesn’t move anything forward. At best, we’ve added an unnecessary bystander to a private difficulty. At worst, we’ve created a trail of secondhand accounts that takes on a life of its own — shaping how people see us and the PI long after the original issue has passed.
Before looping someone in, ask honestly: Can this person actually do something? Do they have the standing to influence the outcome? If the answer is no, keep the circle small, find a mentor or supervisor with real standing to advise, and protect your own reputation by keeping difficult situations from becoming unnecessarily public.
The colleagues who navigate this work most gracefully are not the ones who were never frustrated. They are the ones who stayed curious long enough to find out what was actually going on.
A closing thought: The difficult moments in this work are not detours from our development as biostatisticians — they are part of it. The judgment to know when to push and when to listen, when to advocate and when to understand, builds slowly through experience and reflection. Approach hard situations assuming good faith, pay attention to what you don’t yet know, and remember that the person across the table is usually doing the best they can with what they have.