“Grant” gets used as a catch-all in casual conversation, but research administrators and investigators deal with a structured taxonomy underneath it — funding instruments that differ in legal basis, purpose, selection method, and restriction, each carrying its own eligibility rules, reporting obligations, and administrative burden. Knowing which type of grant a given opportunity actually is changes how a proposal gets built, what a budget can and can’t include, and what compliance obligations follow the award. This guide is the orientation map: it sorts research funding into the major categories researchers and sponsored-programs offices actually encounter, and links out to CASRAI’s more detailed treatment of each one rather than re-deriving it here.
Grant, contract, or cooperative agreement: the instrument itself
Before sorting grants by purpose, it’s worth fixing the broader category they sit in. Under US law, federal awards fall into three legally distinct instrument types, governed by the Federal Grant and Cooperative Agreement Act of 1977 (31 U.S.C. §§ 6303–6305):
- Contract — used when the government is acquiring goods or services for its own direct benefit. Contracts are FAR-governed (Federal Acquisition Regulation), not Uniform Guidance-governed, and generally carry the heaviest procurement and reporting requirements.
- Grant — assistance to carry out a public purpose, where the funding agency anticipates little involvement in how the work is actually conducted. Most investigator-initiated research funding — an NIH R01 or an NSF standard grant, for example — is a grant in this legal sense.
- Cooperative agreement — also assistance toward a public purpose, but the agency anticipates “substantial involvement” in carrying out the funded activity: work-plan approval, steering-committee participation, or active programmatic direction. NIH’s U-series activity codes (U01, U19, U54) and large multi-site or infrastructure awards commonly use this instrument.
“Substantial involvement” isn’t bright-line-defined in the statute; routine progress-report review, no-cost-extension approval, or an audit does not by itself convert a grant into a cooperative agreement. For the full legal test and worked federal-agency examples, see CASRAI’s Grant vs. Contract vs. Cooperative Agreement comparison and the Cooperative Agreement dictionary entry. This guide focuses on the “grant” branch of that split — the categories below are all ways of subdividing grants and grant-like assistance funding further.
Grants by purpose
The most practically useful way to sort grants is by what the money is actually funding, since purpose drives eligibility, review criteria, and post-award obligations more than any other single factor.
Research project grants
The default category most people mean when they say “grant”: funding tied to a specific, defined research project with stated aims, methods, and a budget. The award document restricts the institution to that project — funds generally can’t be redirected to a different research question, even one led by the same investigator, without prior agency approval. NIH’s R01 (Research Project Grant) and NSF’s standard and continuing grants are the canonical examples. See NIH Grants: An Overview and NSF Grants: An Overview for the full mechanism taxonomies each agency runs.
Training and career development grants
A distinct category whose purpose is developing the researcher, not funding a project outcome per se. These carry career-stage and degree-status eligibility rules, review criteria that weight the training environment and mentorship plan alongside the science, and post-award obligations like payback provisions or trainee-outcome reporting that project grants don’t have. In the US, NIH organizes this space through F-series individual fellowships, K-series career development awards (see CASRAI’s Career Development (K) Awards guide and the K99/R00 guide), and T-series institutional training grants that fund training programs rather than individual trainees. Outside the US, UKRI’s Centres for Doctoral Training and the EU’s Marie Skłodowska-Curie Actions Innovative Training Networks fund cohort-based doctoral and early-career training on a comparable model.
Center and infrastructure grants
Funding for shared facilities, multi-investigator centers, or major equipment rather than a single investigator’s project — NIH’s P-series center grants and S10 Shared Instrumentation Grants are examples. These often function administratively more like cooperative agreements because the funder retains active involvement in how the shared resource is governed and used. See CASRAI’s Core Facility (Research Core) guide for how the funding and cost-recovery sides of this category work together.
Small business grants (SBIR/STTR)
A statutorily distinct category under the Small Business Administration’s SBIR/STTR Policy Directive, run in parallel by multiple federal agencies (NIH, NSF, DOE, DOD, and others) under a shared framework but separate agency-specific competitions. Structured in phases — feasibility (Phase I), full R&D (Phase II), and commercialization (Phase III, typically non-SBIR-funded) — with STTR additionally requiring a formal partnership with a nonprofit research institution. See SBIR Phase II and CASRAI’s SBIR/STTR Reauthorization guide.
Grants by restriction: categorical, block, and formula
A second, less commonly discussed axis — borrowed from US fiscal-federalism and public-administration scholarship rather than native research-administration vocabulary, but genuinely useful for understanding why research funding behaves the way it does — sorts grants by how much discretion the recipient has over how the money is spent:
- Categorical grants — restricted to a narrowly defined purpose, with detailed conditions on use. Nearly all federal research funding is categorical in this sense, just categorical at the level of an individual funded project rather than a broad program area: an R01 award is tied to the specific aims and budget in the approved application, full stop.
- Formula grants — a subtype of categorical grant allocated by a statutory formula (population, need, or a similar metric) rather than by competitive review. Uncommon in investigator-initiated research funding, more common in public-health and education contexts.
- Block grants — authorized for a broad policy area, giving the recipient substantial discretion over exactly how the funds are spent within it, with lighter federal reporting than categorical funding. Genuinely block-like funding is rare in the research-grant context; general institutional support or unrestricted indirect-cost recovery are the closest research-administration analogs.
See CASRAI’s Categorical Grant dictionary entry for the full public-administration framing and how it maps onto federal research awards, including allowable-use restrictions, reporting granularity, and limited fungibility across awards — the practical consequences a sponsored-programs office deals with because most research funding sits at the restrictive end of this spectrum.
Grants by selection method: competitive and non-competitive
A third axis, independent of both instrument type and purpose: how the award was selected.
- Competitive grants — awarded through peer review against other applications responding to the same announcement, on the basis of scientific or technical merit. The large majority of investigator-initiated research funding is competitive.
- Non-competitive grants — awarded without a competing pool, typically because a statute, formula, or sole-source justification directs funding to a specific recipient or continues an already-reviewed award (a continuation or renewal that doesn’t re-enter full competitive review, for instance).
See CASRAI’s Competitive Grant and Non-Competitive Grant dictionary entries, and Continuation Award for how an already-funded multi-year project typically continues without re-competing each budget period.
Grants by funder sector
Purpose, restriction, and selection method all cut across a further, practical distinction: who is actually issuing the money.
- US federal agency grants — NIH, NSF, DOE, DOD, and other agencies each maintain their own mechanism taxonomies under the shared Uniform Guidance cost-principles framework (2 CFR 200). See CASRAI’s NIH and NSF overview guides.
- State and institutional grants — state government research funding and internal/institutional seed funding, generally smaller in scale and governed by state or institutional policy rather than federal regulation.
- International and national funding bodies — UKRI and Horizon Europe in the UK/EU, Canada’s Tri-Agency (CIHR, NSERC, SSHRC), Australia’s ARC and NHMRC, and comparable national funding councils elsewhere, each with its own award-instrument conventions. See CASRAI’s Grants Management & Research Funding pillar for the full international coverage.
- Foundation and philanthropic grants — funding from private foundations and nonprofits, which sit outside 2 CFR 200 unless the funds themselves originate from a federal pass-through. These typically distinguish between a project grant (restricted to a specific proposal, much like a categorical federal award) and a general operating grant (unrestricted support for the organization’s mission), a distinction that maps onto the categorical/block axis above even though foundations rarely use that terminology.
Why the taxonomy matters administratively
These four axes — instrument (grant/contract/cooperative agreement), purpose (project/training/center/small-business), restriction (categorical/formula/block), and selection method (competitive/non-competitive) — are independent of each other and combine on any given award. An NIH R01 is a grant, a research project grant, categorical, and competitive, all at once. A large multi-site NIH cooperative agreement is a cooperative agreement, a center/infrastructure award, still categorical, and still competitively selected. Knowing where a specific opportunity sits on all four axes, before writing a proposal or negotiating an award, is what determines the applicable cost principles, the review criteria to write against, the reporting cadence to plan for, and how much programmatic latitude the funded team will actually have once the award is made.
Frequently asked questions
What is the difference between a grant and a contract?
A contract is used when the government is acquiring goods or services for its own direct benefit and is FAR-governed; a grant is assistance toward a public purpose where the funder anticipates little involvement in how the work is carried out, governed instead by Uniform Guidance (2 CFR 200). See the full comparison at Grant vs. Contract vs. Cooperative Agreement.
What is the difference between a grant and a cooperative agreement?
Both are assistance instruments toward a public purpose. The distinction is whether the funding agency anticipates “substantial involvement” in carrying out the work — active programmatic direction, work-plan approval, steering-committee participation — which makes it a cooperative agreement, versus the comparatively hands-off relationship of a standard grant.
What is the difference between a categorical grant and a block grant?
A categorical grant restricts funding to a narrowly defined purpose with detailed use conditions; a block grant is authorized for a broad policy area and gives the recipient substantial discretion over how funds are spent within it. Nearly all federal research funding is categorical, just categorical at the level of an individual project rather than a program area.
What is a formula grant?
A subtype of categorical grant allocated by a statutory formula — population, need, or a comparable metric — rather than through competitive peer review. It’s uncommon in investigator-initiated research funding and more typical of public-health or education program funding.
What is the difference between a project grant and a general operating grant?
This distinction is most common in foundation and philanthropic funding: a project grant restricts funds to a specific proposed activity, while a general operating grant provides unrestricted support for the recipient organization’s broader mission — the same restriction-versus-discretion axis that separates categorical from block funding in the federal context.
