Writing Collaborative and Multi-PI Proposals: How to Structure Teams That Reviewers Trust

Why Collaborative Proposals Exist and Why They Fail

Modern research problems rarely fit neatly within one lab, one discipline, or one institution. Federal agencies recognized this decades ago. NIH introduced its Multiple Principal Investigator (MPI) model in 2007 to formally accommodate leadership structures beyond the single-PI paradigm. NSF has long supported collaborative proposals through both single-submission and separately-submitted pathways. DOE, DOD, USDA, and most other federal agencies now expect multi-institutional teams for their largest awards.

The premise is straightforward: complex problems demand diverse expertise. A translational cancer study may require a molecular biologist, a biostatistician, a clinical oncologist, and a health economist. An NSF Convergence Accelerator project might pair computer scientists with sociologists and community organizations. No single investigator covers that ground alone.

Yet collaborative proposals fail at disproportionate rates. Reviewers consistently flag vague leadership plans, poorly justified team compositions, and governance structures that amount to little more than "we will meet regularly and resolve disagreements collegially." The proposals that succeed treat team structure as a core element of the research design, not an administrative afterthought.

Understanding the Agency Frameworks

NIH Multiple PD/PI Model

NIH allows any number of PD/PIs on a research project grant. The requirements are specific:

Contact PD/PI. One PD/PI must be designated as the Contact PD/PI. This person serves as the primary point of communication with NIH and must be affiliated with the applicant organization. Critically, the Contact PD/PI holds no special authority within the team — the role is administrative, not hierarchical. Reviewers who see the Contact PI described as the "lead" or "senior" PI will question whether the MPI structure is genuine or cosmetic.

Leadership Plan. Every MPI application must include a Multiple PD/PI Leadership Plan as a separate attachment. The plan must address:

• Rationale for the multi-PI approach
• Governance and organizational structure
• Communication plans
• Decision-making processes for scientific direction
• Procedures for resolving conflicts
• Roles, responsibilities, and areas of expertise for each PD/PI

eRA Commons registration. All PD/PIs must hold eRA Commons accounts with PI role prior to submission. This is a hard compliance requirement — missing it will trigger an administrative return.

Budget considerations. When PD/PIs are at different institutions, standard consortium/subaward budget instructions apply. NIH does not require a per-PI budget breakdown in the application itself, but the Leadership Plan should address resource allocation, and many institutions require internal budget splits to be documented before submission.

NIH Simplified Peer Review Framework (2025)

For applications submitted on or after January 25, 2025, NIH reorganized its five review criteria into three factors:

• Factor 1: Importance of the Research (Significance + Innovation) — scored 1-9
• Factor 2: Rigor and Feasibility (Approach) — scored 1-9
• Factor 3: Expertise and Resources (Investigators + Environment) — evaluated as "appropriate" or "additional expertise/resources needed"

Factor 3 is where team structure gets evaluated. Reviewers assess whether investigators have the background, training, and expertise for the proposed work, and — for MPI applications specifically — the quality of the leadership plan to facilitate coordination and collaboration. An "additional expertise/resources needed" determination under Factor 3 can sink an otherwise strong application, because it signals that the review panel doubts the team can execute the science.

NSF Collaborative Proposals

NSF offers two submission pathways for multi-institutional collaborations:

Single submission with subawards. One lead organization submits the entire proposal and includes subaward budgets for partner institutions. The lead PI bears primary responsibility for all award management. This pathway gives the lead organization more control but creates administrative burden and can delay fund access for subawardee institutions.

Separately submitted collaborative proposals. Each participating organization submits its own proposal through Research.gov. All proposals must share the same title, prefixed with "Collaborative Research:" and NSF links them for joint review. If funded, each institution receives a separate award. NSF generally considers this the preferred pathway because it distributes administrative responsibility and gives each institution direct access to its funds.

For separately submitted proposals, all linked proposals must be submitted within a reasonable window — most sponsored programs offices recommend 48 hours or less to avoid complications.

NSF evaluates collaborative proposals using its standard criteria (Intellectual Merit and Broader Impacts) but reviewers pay close attention to whether the collaboration is genuine — whether the project requires the combined expertise of the team, or whether it is really a collection of loosely related projects stapled together.

Building a Team That Reviewers Believe In

Start with the Science, Not the Roster

The most common structural mistake in collaborative proposals is assembling a team before defining the problem. Reviewers can immediately tell when a team was built around existing relationships rather than around the demands of the research question. The proposal reads like a roster of friends who decided to find something to work on together, rather than a group of experts whose combined skills are uniquely suited to a specific challenge.

Begin with the research aims. For each aim, identify the methodological expertise required. Where do those competencies overlap, and where are there gaps that no single investigator can fill? Those gaps define where your collaborators need to come from. When you write the Leadership Plan, this logic should be transparent: "Aim 2 requires longitudinal analysis of electronic health records across three health systems. Dr. Vasquez's group at [Institution B] maintains validated EHR data pipelines for these systems and has published extensively on bias correction methods for multi-site clinical data. No comparable resource exists at the lead institution."

Justify Every Team Member

Every PI and co-investigator named on the proposal should have a clear, non-overlapping role tied to specific aims or project components. Reviewers are skeptical of large teams where multiple investigators appear to cover the same territory. If two PIs both list "data analysis" as their primary contribution, the reviewer's question is immediate: why do you need both?

Effective role descriptions follow a pattern:

• What this person contributes (specific expertise, methods, resources, or access)
• Why this contribution is essential (what would be missing without it)
• Where in the project plan this contribution appears (linked to specific aims, tasks, or milestones)

For junior investigators on an MPI application, reviewers will scrutinize whether they have genuine scientific autonomy or are functionally serving as senior personnel under a PI title. The time commitments should reflect meaningful leadership responsibility — a PD/PI listed at 5% effort on a five-year R01 does not look like a principal investigator.

Demonstrate Prior Collaboration (But Do Not Rely on It)

Citing prior joint publications, shared grants, or co-mentored trainees is useful evidence that a team can work together. But reviewers distinguish between "we have successfully collaborated before" and "we have the structures in place to manage this specific project." A track record of collaboration is necessary but not sufficient. The Leadership Plan must still lay out concrete governance for the proposed work, not gesture at past experience as a substitute.

When prior collaboration does not exist — as is often the case when agencies specifically seek new interdisciplinary partnerships — you need to be candid about it. Describe what structures you are putting in place precisely because this is a new collaboration: formalized communication schedules, defined escalation paths, and shared project management infrastructure. Reviewers respect honesty about the challenge of building a new team far more than they respect vague assurances from investigators who have never worked together claiming they will "leverage complementary expertise."

Writing the Leadership Plan

The Leadership Plan is the single document where reviewers assess whether your team can actually function as described. Under the NIH Simplified Review Framework, it directly informs the Factor 3 evaluation. Weak leadership plans are one of the most commonly cited concerns in summary statements for collaborative applications.

Rationale

Open with a concise statement of why the multi-PI structure is essential for the proposed research. This is not a general argument for team science. It is a specific argument that this project requires the combined authority, expertise, and resources of the named PD/PIs, and that a single-PI model with co-investigators would be inadequate.

A strong rationale might read: "This project integrates two distinct methodological traditions — small-molecule screening (Dr. Okafor) and patient-derived organoid modeling (Dr. Lindgren) — each requiring independent laboratory infrastructure, personnel management, and scientific decision-making. Neither approach is subordinate to the other; both are co-equal pillars of the research design. A single-PI structure would misrepresent the scientific contributions of one investigator as ancillary support rather than intellectual leadership."

A weak rationale reads: "The complexity and scope of this project require a team-based approach. Drs. Smith and Jones have a long history of productive collaboration and bring complementary expertise to this work."

Organizational Structure

Describe the governance hierarchy explicitly. Most successful leadership plans include:

Steering committee or executive committee. Composed of all PD/PIs, meeting at defined intervals (monthly or quarterly depending on project scope). Responsible for strategic decisions about scientific direction, resource reallocation, and milestone assessment.

Project manager or coordinator. Including a dedicated project manager signals to reviewers that you understand the administrative demands of a multi-site project. Experienced reviewers have noted that omitting a project manager is a red flag — it suggests the PIs do not appreciate the coordination burden.

Working groups or cores. For larger projects, describe how sub-teams are organized around specific aims or functional areas, who leads each, and how they report to the steering committee.

An organizational chart is valuable for complex teams. NIH explicitly encourages it. Keep it clean — boxes and lines showing reporting relationships, with PD/PI names, institutional affiliations, and aim assignments visible at a glance.

Roles and Responsibilities

Dedicate a paragraph to each PD/PI specifying:

• Scientific leadership responsibilities (which aims, cores, or project components they direct)
• Administrative responsibilities (budget management for their site, personnel supervision, IRB/IACUC oversight)
• Percent effort and how it is distributed across project activities
• Unique qualifications that make them essential to the team

Avoid the temptation to describe all PIs as jointly responsible for everything. Reviewers view this as unrealistic and evasive. Even on tightly integrated projects, each PI should have defined areas of primary responsibility.

Communication Plan

Specify the communication infrastructure with enough detail to be credible:

• Weekly or biweekly PI calls (platform, standing agenda structure, who takes notes)
• Monthly all-hands meetings for full project team
• Annual in-person retreats (budgeted as a travel line item)
• Shared project management tools (mention the platform — REDCap for data, Slack or Teams for asynchronous communication, a shared document repository)
• Reporting cadence between working groups and the steering committee

Vague statements like "the PIs will communicate regularly" earn no credit with reviewers. Specificity demonstrates that you have thought through the operational reality of managing a distributed team.

Decision-Making Process

Describe how scientific decisions are made. A common framework:

1. Working-level decisions (day-to-day experimental adjustments, personnel scheduling) are made by the responsible PI within their area of authority.
2. Cross-cutting decisions (changes to shared protocols, timeline adjustments affecting multiple sites) are brought to the steering committee for discussion and consensus.
3. Strategic decisions (redirecting aims based on interim results, requesting a no-cost extension, dropping or adding a sub-study) require a formal steering committee vote, with the Contact PI facilitating but not holding veto authority.

Define what happens when consensus cannot be reached. A typical escalation path involves consultation with the program officer and, for irreconcilable disputes, mediation by a named external advisor — often a senior scientist outside the project who is respected by all parties and has agreed in advance to serve in this role.

Conflict Resolution

Reviewers specifically look for conflict resolution procedures, and the best plans go beyond the generic. Address:

• Scientific disagreements: How are disputes over methodology, data interpretation, or analytical approaches resolved? Voting? Mediation? Binding arbitration by an external advisory board?
• Resource conflicts: What happens if one site needs more funds than originally allocated? Who has authority to approve budget reallocations, and up to what threshold before NIH approval is required?
• Personnel disputes: What is the process if a co-investigator at one site is underperforming or if a PD/PI needs to reduce effort?
• Departure of a PD/PI: What happens if a PD/PI leaves the project or changes institutions? Who assumes their responsibilities? Is there a succession plan?

A concrete example: "In the event of a scientific disagreement that cannot be resolved by the PD/PIs, the dispute will be referred to Dr. [Name], Professor of [Field] at [Institution], who has agreed to serve as an external mediator. Dr. [Name] is not involved in the proposed research and has no financial relationship with either institution. The mediator's recommendation will be binding for a period of 12 months, after which the PD/PIs may revisit the decision."

Budget Strategy for Multi-Site Teams

Allocating Funds Across Institutions

Budget allocation in collaborative proposals must reflect the actual distribution of work. Reviewers compare the budget splits against the described roles and responsibilities. A proposal where Institution A receives 80% of the funds while Institution B's PI is described as a co-equal leader will raise immediate questions.

For NIH applications with consortium/subaward arrangements, the lead institution includes a subaward line item in its budget, and detailed budgets for each subaward institution are included as separate budget pages. The 8% de minimis indirect cost rate or a negotiated rate applies to subaward costs (note: only the first $25,000 of each subaward counts toward the lead institution's modified total direct costs for F&A calculation purposes).

For NSF separately submitted collaborative proposals, each institution submits its own budget directly to NSF. This simplifies the indirect cost issue since each institution uses its own negotiated rate, but it requires careful coordination to ensure the budgets are internally consistent — if Institution A's budget assumes 200 hours of sequencing services from Institution B, Institution B's budget must include matching capacity.

Common Budget Pitfalls in Collaborative Proposals

Unfunded collaboration. Naming a PI at a partner institution without any budget allocation to that institution signals a paper collaboration. If someone is important enough to be a PD/PI, their institution should receive funds to support their participation.

Duplicate resources. Requesting the same equipment at two sites when the project plan only requires one site to perform that function. Reviewers will catch this.

Missing coordination costs. Travel for in-person meetings, shared software licenses, project management personnel — these are legitimate and expected costs in a multi-site project. Omitting them suggests the team has not realistically planned for the coordination overhead.

Mismatched effort and responsibility. A PI described as leading two aims but budgeted at 10% effort, while a co-investigator with a narrower role is budgeted at 25%. The effort percentages should tell the same story as the role descriptions.

Budget Justification Language

For multi-site proposals, the budget justification must explain subaward costs with the same rigor applied to internal costs. For each consortium partner, describe:

• The specific scope of work to be performed at that institution
• Why the work must be performed there (unique equipment, patient population access, regulatory approvals, specialized expertise)
• Key personnel at that site and their roles
• Major cost categories and their justification

Generic statements like "subaward to [University] for collaborative research activities" are insufficient. Treat each subaward justification as a mini-proposal within the budget narrative.

Strategies That Strengthen Reviewer Confidence

Include a Collaboration Timeline

A Gantt chart or milestone table that shows which institution is responsible for each deliverable, when handoffs between sites occur, and where the critical dependencies lie. This demonstrates that you have mapped the project's logistics, not just its science. Place it in the Research Strategy or as a supplementary figure if page limits allow.

Name Your External Advisory Board Early

If the project warrants an external advisory board — and most large collaborative projects do — name the members in the proposal. An advisory board composed of "TBD" members communicates that the governance structure is aspirational rather than operational. Named advisors who have provided letters of support carry far more weight.

Address Data Sharing and Intellectual Property

Multi-institutional collaborations generate questions about data ownership, publication rights, and IP. Reviewers notice when these issues are ignored. The Leadership Plan should include at least brief statements on:

• How data will be shared among sites during the project (shared repositories, transfer protocols, access controls)
• Publication policies (who has authorship rights, how disputes about authorship are handled, whether any site has veto power over publication)
• IP ownership framework (typically governed by the Bayh-Dole Act and each institution's IP policies, but worth acknowledging)

Write Integrated, Not Parallel, Aims

The strongest collaborative proposals have aims that are genuinely interdependent — where the output of Aim 1 at Institution A is the input to Aim 2 at Institution B. Proposals where each PI essentially runs an independent project under a shared title are flagged as "collaborative in name only." Reviewers test this by asking: if one site's work were removed, would the remaining aims still stand? If yes, the collaboration is not well-motivated.

Tailor the Team Size to the Problem

Larger teams are not inherently stronger. A two-PI collaboration with clear complementarity and tight integration will outscore a five-PI team with diffuse responsibilities and unclear governance. Add collaborators only when the science demands it. Every additional PI increases the coordination burden and gives reviewers more opportunities to question whether the team is manageable.

Navigating Agency-Specific Requirements

NIH: Letters of Support vs. Consortium Agreements

Letters of support from key collaborators confirm their commitment and specify what they will contribute. These are distinct from consortium agreements (formalized inter-institutional contracts executed after award). Reviewers want to see letters from all consortium PIs and from any individual providing critical resources or access. Each letter should reference the specific project, not read as a generic endorsement.

NSF: Facilities, Equipment, and Other Resources

NSF requires a Facilities, Equipment, and Other Resources document for each collaborating institution. This is where you describe the institutional infrastructure supporting the work — lab space, computing resources, shared instrumentation, specialized facilities. For collaborative proposals, this document should also describe any shared infrastructure between institutions (joint data centers, reciprocal facility access agreements).

DOD and DOE: Teaming Arrangements

For Department of Defense BAAs and DOE FOAs, collaborative teams often form through formal teaming arrangements before proposal submission. The prime contractor (usually the lead institution or company) submits a single proposal with subcontract budgets. These agencies tend to scrutinize management plans with particular attention to milestone-driven deliverables, go/no-go decision points, and government oversight mechanisms. Management plans for DOD and DOE proposals are typically more detailed and more formal than NIH leadership plans.

Frequently Asked Questions

What is the difference between a co-PI, a co-investigator, and a multiple PD/PI?

The terminology varies by agency and institution. At NIH, a PD/PI (Program Director/Principal Investigator) holds primary responsibility for the scientific and technical direction of the project. Under the MPI model, multiple individuals share this responsibility equally. A co-investigator contributes to the scientific development or execution of the project but does not bear the same level of responsibility as a PD/PI. NIH does not use the term "co-PI" — that is an institutional designation some universities apply for internal purposes. At NSF, a co-PI is formally recognized and may appear on up to three pending or awarded proposals at a time (compared to unlimited for senior personnel). Understanding these distinctions matters because the title you assign to each team member carries specific implications for responsibility, authority, and compliance.

How long should the NIH Multiple PD/PI Leadership Plan be?

NIH does not specify a page limit for the Leadership Plan. In practice, effective plans for two-PI applications run one to two pages. Larger teams with three or more PD/PIs or complex multi-site structures may require two to three pages. The plan should be comprehensive but not padded. Every paragraph should address a required element — rationale, governance, communication, decision-making, conflict resolution, or roles and responsibilities. Reviewers will skim a plan that reads like boilerplate and focus on whether the substantive requirements are addressed with specificity.

Should we use the MPI model or a single PI with co-investigators?

Use the MPI model when two or more investigators make genuinely co-equal intellectual contributions that cannot be accurately represented in a single-PI hierarchy. If one investigator is clearly leading the project and others are contributing within their defined scope, a single PI with co-investigators is the appropriate and simpler structure. Do not use MPI as a courtesy appointment or to satisfy a collaborator's desire for PI status — reviewers will see through it, and the added administrative complexity of the Leadership Plan becomes a liability rather than an asset. The MPI model is strongest when both PIs bring independent lines of inquiry that converge on the proposed research question.

How do we handle budget allocation when one institution is doing significantly more work?

Unequal budget splits are normal and expected. The allocation should match the distribution of effort, personnel, and resources described in the Research Strategy and Leadership Plan. A 70/30 or 80/20 split is perfectly defensible when the scope of work justifies it. What reviewers flag is inconsistency — a budget that tells a different story than the narrative. If Institution B receives 20% of the budget, the PI at Institution B should not be described as leading half the aims. Acknowledge the asymmetry in the Leadership Plan and explain it: "Dr. Nakamura's contribution is concentrated in Years 1-2 during the assay development phase, with reduced effort in Years 3-5 as the project shifts to clinical implementation led by Dr. Fernandez."

What happens if a PD/PI leaves the project mid-award?

NIH requires prior approval for changes in PD/PI status. If a PD/PI withdraws, the remaining PD/PIs and the grantee institution must notify the awarding institute and propose a plan — either adding a replacement PD/PI, restructuring the leadership, or converting to a single-PI award. The best Leadership Plans address this contingency proactively, naming potential successors or describing the process for identifying a replacement. For NSF collaborative awards where each institution holds a separate grant, the departure of one PI may require a formal amendment to the remaining awards. Building succession language into the original Leadership Plan demonstrates forethought and reduces the administrative disruption if a transition becomes necessary.

---

Collaborative proposals demand that you treat team design with the same rigor you bring to experimental design. The agencies have created formal structures — MPI models, collaborative submission pathways, leadership plan requirements — because they recognize that team science is hard to manage and easy to do poorly. Granted helps teams draft leadership plans, align budgets across institutions, and structure governance frameworks that satisfy reviewer expectations from the first submission.