Universities move materials, data, confidential information, and funded research obligations across institutional boundaries constantly — and each of those movements is governed by a different legal instrument, negotiated and signed by a different office, with a different realistic timeline. Picking the wrong instrument, or letting the wrong person sign it, is one of the most common and most expensive process mistakes in university research administration. This guide is the decision layer that sits above CASRAI’s instrument-specific deep dives: it maps the situation to the agreement, explains who actually has authority to sign, and covers the clauses — publication rights above all — where academic and industry expectations collide most often.
For the mechanics of any single instrument, see CASRAI’s dedicated guides: the MTA process guide (UBMTA, non-standard MTAs, sticking points) and publication delay and review clauses in sponsored research agreements. This page is the map that tells you which of those — or which of several other instruments — applies to your situation in the first place.
The decision table: what are you moving, and what does it trigger?
The starting question is always the same: what is actually changing hands — tangible material, data, confidential information, or a piece of funded research work? The table below maps the common scenarios research offices see to the agreement type, who typically signs it, a realistic turnaround, and the single clause category most likely to stall the negotiation.
| What you’re transferring or receiving | Agreement type | Who signs | Typical turnaround | Clause most likely to stall it |
|---|---|---|---|---|
| Tangible research material — cell line, plasmid, antibody, reagent, tissue sample, animal model | Material Transfer Agreement (MTA) | Institution (TTO / research contracts office), not the PI | Days–weeks if a UBMTA signatory template applies; weeks–months for a non-standard/industry MTA | Reach-through rights on future discoveries; ownership of modifications/progeny |
| Restricted or identifiable dataset from a single source, one-directional access | Data Use Agreement (DUA) | Institution (sponsored programs / privacy office) | Days–weeks for a standard template; longer if re-identification risk or HIPAA/FERPA data is involved | Permitted secondary uses; re-identification and security safeguard requirements |
| Bidirectional or multilateral data exchange between collaborating institutions | Data Sharing Agreement (DSA) | Institution (sponsored programs / research contracts office) | Weeks, longer for multi-party agreements | Downstream sharing and publication rights over the combined dataset |
| Confidential information only — no material, no funding, evaluating a possible collaboration | Confidentiality Agreement / NDA | Institution (TTO or research contracts office); mutual NDAs are usually fast to execute | Days–1–2 weeks with a standard institutional template | Overbroad definition of “confidential” that would restrict the researcher’s own prior or independent work |
| One institution funds another institution to carry out a defined piece of collaborative research, outside a federal prime award | Sponsored Research Agreement (SRA) / research collaboration agreement | Institution (sponsored programs director, VP research, or delegated contracts officer) | Weeks for a repeat industry partner using a standard template; 2–6 months for a first-time partner or non-standard terms | Publication rights and delay; IP ownership of resulting inventions; indemnification |
| A defined share of a federal (or federal flow-through) award’s programmatic scope and budget is delegated to a collaborating institution that will exercise its own judgment in carrying out part of the research | Subaward | Institution (sponsored programs office); subrecipient’s institutional signatory countersigns | Days with a Federal Demonstration Partnership (FDP) template between FDP-member institutions; weeks–months for first-time, non-FDP, or foreign subrecipients | Flow-down of the prime award’s compliance terms; F&A rate treatment; export control on foreign subawards |
| A company or vendor delivers a defined product or service (e.g., assay runs, sequencing, custom fabrication) under the institution’s direction, without exercising independent programmatic judgment | Subcontract / purchase order (procurement) | Institution’s procurement or contracts office | Days–weeks; governed by the institution’s standard procurement process, not sponsored-programs review | Scope/deliverable specificity; payment terms tied to acceptance criteria |
| An individual provides discrete expert advice or a bounded service, with no delegated programmatic decision-making authority | Consultant agreement | Institution (department or sponsored programs office, depending on dollar threshold); sometimes department-level signature authority for small amounts | Days–1–2 weeks with a standard consulting template | Conflict-of-interest and effort disclosure; IP assignment for anything the consultant creates |
Two situations frequently need more than one instrument at once: a collaboration that involves both a physical shipment and an accompanying dataset needs an MTA and a DSA or DUA — neither substitutes for the other. And an SRA or subaward that will involve the exchange of proprietary background information ahead of execution often needs a short NDA to cover the negotiation period itself, before the substantive agreement is signed.
Why a principal investigator cannot sign a research agreement
This is the single most common and most costly process mistake in this area, and it isn’t a formality — it’s a real legal distinction. A PI is an employee of the institution, not an agent empowered to bind it to a contract. Signature authority — the legal capacity to commit the institution to the agreement’s obligations, indemnities, and financial terms — is vested in a specific, delegated set of institutional officials: typically the sponsored programs director, a research contracts officer, the vice president or vice provost for research, or general counsel, depending on the agreement’s dollar value and risk profile. This is the same pattern used across every instrument in the table above and mirrors the “Authorized Organizational Representative” (AOR) role institutions designate for federal grant and contract submissions.
The practical reasons this matters:
- Enforceability. An agreement signed by someone without delegated authority may not bind the institution at all — the sponsor could find itself with an agreement that isn’t actually enforceable against the university, which defeats the point of having a written contract.
- Institutional risk exposure. Indemnification, liability, IP ownership, and governing-law terms carry financial and legal consequences for the whole institution, not just the lab. Only the office with authority to accept that risk on the institution’s behalf should be the one agreeing to it.
- Compliance flow-down. Federal and institutional compliance terms (cost principles, export control, data security, human-subjects and biosafety requirements) need to be checked and flowed down correctly, which is exactly the review a PI is not positioned or authorized to perform.
- The cleanup cost. When a PI does sign — often because a sponsor’s counterpart pushed for a fast signature — the usual outcome is either a delay while the institution decides whether to ratify the agreement after the fact, or a period where the parties are operating on a document the institution doesn’t consider binding. Both outcomes cost more time than routing the agreement through the contracts office in the first place would have.
The PI’s real role is to identify the need, describe the scope of work or material, and support the negotiation on the technical substance — not to execute the document. Research offices exist specifically to carry that signature authority so PIs don’t have to, and so the institution’s risk position is applied consistently across every agreement it enters.
Subaward vs. subcontract vs. consultant: the determination that gets missed
These three terms get used loosely and interchangeably in practice, but they are legally distinct categories with different compliance obligations attached — and getting the classification wrong on a federally funded project is a real audit finding, not just a paperwork nuance.
For work funded under a federal award, the controlling question is the subrecipient vs. contractor determination under the Uniform Guidance (2 CFR 200.331). The regulation sets out characteristics of each, applied in substance rather than by the label on the document:
- Subrecipient (subaward) characteristics: determines who is eligible to receive federal assistance; has performance measured against whether the objectives of the federal program were met; has responsibility for programmatic decision-making; has responsibility for adherence to applicable federal program requirements; and uses the federal funds to carry out a program of the organization, rather than to provide goods or services for the pass-through entity’s own use.
- Contractor (procurement relationship — a subcontract or purchase order) characteristics: provides the goods or services within normal business operations; provides similar goods or services to many different purchasers; operates in a competitive environment; provides goods or services that are ancillary to the operation of the federal program; and is not subject to compliance requirements of the federal program as a result of the agreement (though it may be subject to normal procurement or contractual requirements).
No single characteristic is decisive on its own — the determination is made on the substance of the relationship as a whole, and the same regulation expects the pass-through entity to make and document that judgment before the award goes out, not after. See CASRAI’s subaward agreement negotiation guide for how the subaward instrument itself is structured and negotiated once that determination is made.
A consultant agreement sits alongside this distinction rather than inside it: it covers a discrete engagement with an individual (not another institution) who is providing expert advice or a bounded service without exercising independent programmatic judgment over any part of the funded project’s scope. Consultant relationships raise their own compliance questions — conflict-of-interest and effort disclosure, and IP assignment for anything the consultant creates in the course of the engagement — but they are not evaluated against the subrecipient/contractor test, since a consultant is an individual providing a service, not an organization potentially carrying out part of the program.
Publication rights: where academic and industry expectations collide
This is the clause category that stalls more sponsored research and collaboration negotiations than almost any other, because it sits directly on top of a difference in institutional mission that isn’t fully negotiable on either side.
A university generally cannot accept an indefinite publication embargo, or a sponsor’s unilateral right to block publication outright. Academic freedom, tenure and promotion review (which depends on a demonstrable publication record), graduate students’ need to publish and defend a thesis, and — for federally funded basic research — the government’s own expectation that funded results be disseminated, all make an open-ended restriction on when or whether results can be published incompatible with the institution’s core function. This is a firm institutional position at most research universities, not a negotiating opening.
What institutions typically can accept, and what sponsors typically ask for instead, is a bounded review-and-delay period:
- A review window — commonly in the range of 30 to 60 days — during which the sponsor may review a manuscript before submission or public disclosure, primarily to identify the sponsor’s own confidential information that should be removed and to flag anything that might affect a pending patent filing.
- An additional delay, sometimes layered on top of the review window, specifically to allow time to file a patent application on a discovery the manuscript would otherwise disclose — often capped so the combined review-plus-delay period doesn’t exceed roughly 90 days in total, though the exact ceiling is negotiated per agreement and varies by institutional policy.
- Redaction, not veto — the sponsor can typically request that its own confidential information be removed from the manuscript, but cannot block publication of the underlying research results once the review/delay period has run.
The specific numbers are negotiated, not fixed by any single external standard, and vary by institution and by sponsor — see CASRAI’s dedicated guide to publication delay and review clauses for a fuller treatment of how these terms are typically structured and negotiated. The point to plan around here is simpler: if a sponsor’s draft agreement proposes an open-ended right to prevent publication, that is very likely to be a hard stop for the university’s research contracts office, and raising it early — before a collaboration’s planned start date — avoids losing weeks to a term that was never going to be acceptable as drafted.
Other clauses that routinely stall negotiations
- IP ownership of improvements. Who owns inventions made using licensed background IP, shared material, or joint research — the provider/sponsor’s pre-existing IP, versus new IP the recipient institution’s researchers create using it. Institutions typically retain ownership of inventions made by their own employees (consistent with the Bayh-Dole framework for federally funded work) while granting the sponsor negotiated rights to license the results; sponsors often want broader ownership or automatic license rights over anything built on their contribution. See CASRAI’s guide on indemnification in university patent license agreements for how the adjacent risk-allocation question is typically handled once licensing terms are on the table.
- Indemnification. Sponsors and material providers frequently request broad indemnification from the recipient institution. Public and private research universities typically resist open-ended indemnification — many public institutions face sovereign-immunity or state-law constraints that limit what they can legally agree to — and negotiate the term down to the institution’s own negligence or misuse rather than a blanket assumption of the sponsor’s risk.
- Governing law and dispute resolution. A sponsor based in one state or country often proposes its own governing law and home-court jurisdiction. Public universities are frequently constrained (by state law or institutional policy) to insist on the law and courts of their own state, and some cannot agree to binding arbitration or a waiver of sovereign immunity at all — a mismatch that has to be resolved before signature, not worked around afterward.
- Onward transfer and export control. Whether material, data, or technical information covered by the agreement can be passed to a third party, and whether the transfer itself triggers export-control review (Export Administration Regulations or International Traffic in Arms Regulations screening), particularly for transfers involving a foreign institution or a foreign national researcher.
Realistic negotiation timelines
Timelines vary by institution, but a consistent pattern holds across research-contracting practice:
- Standard template, no first-time terms — a mutual NDA on an institutional template, a UBMTA Implementing/Simple Letter Agreement, or a subaward between two Federal Demonstration Partnership member institutions — is usually the fastest path: days to a few weeks.
- First-time industry partner, or non-standard terms — a new sponsor’s own SRA template, a non-standard MTA involving patented material, or a subaward to a first-time or foreign subrecipient — commonly takes several weeks to a few months, largely driven by the clauses above (publication rights, IP ownership, indemnification, governing law).
- Complex or unresolved terms — negotiations that stall on reach-through rights, an unacceptable publication restriction, or a governing-law/sovereign-immunity conflict can extend well beyond that, sometimes months, if the parties don’t converge early on which terms are genuinely non-negotiable for the institution.
Build in lead time before a collaboration’s intended start date, and route the agreement to the research contracts or sponsored programs office as soon as the need is identified — not after the sponsor is already asking when they can ship material or release funds. Starting the classification and drafting process early is the single biggest lever a PI actually controls over how long this takes.
Frequently asked questions
Can a principal investigator sign a material transfer agreement, NDA, or sponsored research agreement?
No. These are institutional agreements — the institution, not the individual researcher, is the contracting party and bears the resulting legal and financial obligations. A PI identifies the need and supports the negotiation on technical substance, but a delegated institutional official (sponsored programs director, research contracts officer, VP for research, or general counsel, depending on the agreement) holds the actual authority to sign.
What’s the difference between a subaward, a subcontract, and a consultant agreement?
A subaward goes to an organization that exercises independent programmatic judgment over part of a federally funded project’s scope and is measured against the program’s objectives — it’s a subrecipient relationship under 2 CFR 200.331. A subcontract (or purchase order) goes to a vendor delivering a defined good or service within its normal business operations, without programmatic decision-making authority — a procurement relationship. A consultant agreement covers a discrete engagement with an individual providing expert advice or a bounded service, evaluated separately from the subrecipient/contractor test since a consultant isn’t an organization potentially carrying out part of the program.
Can a sponsor require an indefinite publication delay?
Generally no. Most research universities cannot accept an open-ended restriction on publication because of academic freedom, tenure/promotion requirements, and — for federally funded basic research — the funding agency’s own dissemination expectations. What’s typically negotiable instead is a bounded review period (commonly around 30–60 days) plus, in some cases, an additional delay to allow a patent filing, often with a combined cap around 90 days.
What’s the difference between an MTA and a data use agreement?
An MTA governs tangible research materials — cell lines, reagents, biological specimens, physical samples. A Data Use Agreement governs one-directional access to a dataset, typically a restricted or identifiable one from a single source. If a collaboration involves both a physical shipment and an accompanying dataset, both agreements are usually needed.
Who at a university actually negotiates these agreements?
It depends on the instrument, but the negotiating and signing office is consistently the institution’s research contracts office, sponsored programs office, or technology transfer office (for IP-heavy instruments like MTAs and licenses), sometimes with general counsel involved directly for higher-risk terms like indemnification and governing law. Procurement-type subcontracts and purchase orders typically route through the institution’s procurement office instead.
For the broader technology-transfer landscape this fits into, see CASRAI’s Technology Transfer cluster hub; for the funding and award-management side of these agreements, see the Grants Management cluster hub.
Last verified: August 2026. This guide reflects the Uniform Guidance subrecipient/contractor determination (2 CFR 200.331) and general research-contracting practice at US research universities; specific dollar thresholds, review-period lengths, and signature-authority delegations vary by institution and should be confirmed against your own institution’s policy.
