Research Grants vs Scientific Salaries: Should We Fund Projects or People?

Getting your Trinity Audio player ready...

Research grants finance defined activities; scientific salaries pay researchers for their work. The two overlap because grants often fund salaries. The important policy choice is how much a scientist’s livelihood should depend on repeatedly winning approval for a particular project.

Project grants are useful for budgeting experiments, equipment, and coordinated work. Stable, flexible support for researchers can protect continuity and allow them to follow unexpected discoveries. A strong funding system can combine both.

What is the difference between research grants and scientific salaries?

A research grant is an award governed by a funder’s terms. It may support a specific project, a broader research programme, or an individual fellowship.

A scientific salary is regular employment compensation. Its funding may come from an institution’s budget, an endowment, a grant, or several sources. For example, UKRI’s EPSRC guidance explicitly allows project staff payroll costs.

Consequently, receiving a salary does not necessarily provide security or scientific independence. A researcher’s salary may disappear when a short grant ends; a fellowship may provide years of flexible support.

The comparison below concerns project-dependent funding versus continuing support for researchers, rather than two mutually exclusive accounting categories.

Dimension	Project-dependent grants	Stable support for researchers
Main funding decision	Which proposed work should receive resources?	Whose continuing research should receive support?
Research direction	Usually tied to an approved scope, with varying flexibility	Can allow broader changes of direction
Continuity	Exposed to project end dates and renewal decisions	Stronger when support is committed for several years
Accountability	Spending, research progress, and agreed activities	Quality and integrity of a body of work
Main strength	Allocating resources to defined needs	Preserving expertise and freedom to explore
Main risk	Repeated applications and pressure to promise results	Entrenchment, weak review, or exclusion of newcomers

These are design tendencies. A flexible grant can outperform a restrictive salaried position on research freedom.

Why project grants remain useful

Some research has substantial costs beyond the researcher’s time. A laboratory needs equipment and consumables; a field expedition needs transport and logistics; a large collaboration needs shared infrastructure.

Project grants let funders assess whether the proposed methods, team, facilities, and budget fit together. They also make it possible to direct resources toward an explicit public objective.

Proposal preparation can have value: it can expose an unrealistic budget, an inadequate method, or a missing collaborator. The problem arises when the burden of securing funding becomes disproportionate to what the selection process achieves.

Paying the scientist and financing the experiment are different needs. Covering one does not automatically cover the other.

Where grant dependence can weaken research

Applications consume research time

A 2013 observational study of applicants to Australia’s NHMRC Project Grants scheme estimated an average of 34 working days of researcher time per proposal. This included contributions across researchers, rather than necessarily 34 days from one applicant. The estimate concerns a particular scheme and funding round, not every grant system. Nevertheless, it illustrates a substantial opportunity cost. Herbert and colleagues, BMJ Open.

Longer funding commitments could reduce repeated applications. However, salary-based systems still incur hiring, evaluation, management, and appeal costs. Their efficiency should be measured across the whole system.

Proposals can reward predictability

A project application asks researchers to explain what they intend to do. That is reasonable for planning resources, but the most valuable result may be one nobody could describe beforehand.

If funding renewal depends heavily on delivering the original plan, researchers have a reason to favour predictable work or avoid an unexpected direction. This is an incentive risk, not a claim that every grant discourages discovery.

Funding gaps can interrupt useful work

When employment depends on a particular award, a failed renewal can interrupt a valuable research programme even without evidence that its work has become poor. Researcher support spanning several projects could preserve skills, unfinished investigations, datasets, and software maintenance.

This benefit requires credible funding commitments. A payment labelled “salary” provides little protection if it is unpredictable or too small to support meaningful research time.

What evidence supports funding people?

The Howard Hughes Medical Institute offers an established example. Its Investigator Program supports researchers over renewable seven-year terms and permits substantial freedom to pursue their scientific interests.

A peer-reviewed study by Pierre Azoulay, Joshua Graff Zivin, and Gustavo Manso compared HHMI investigators with similarly accomplished NIH-funded scientists. It found a higher rate of high-impact papers among HHMI investigators and changes in research direction consistent with greater exploration. Incentives and Creativity: Evidence from the Academic Life Sciences (2011).

This supports taking funding duration, flexibility, and tolerance of early failure seriously. It does not isolate the effect of salaries: HHMI combines several features, including selection and research resources. Nor does this comparison establish that the same results would follow across every discipline.

The defensible inference is that how support is structured can influence scientific exploration, not that one payment category universally produces better science.

Scientific salaries need fair accountability

Stable support can create its own problems. An institution may protect established insiders while excluding capable newcomers. An evaluation system may reward reputation rather than useful contributions.

A credible model should therefore include:

• Substantive review: assess methods, findings, software, datasets, and other contributions, rather than relying on publication counts.
• Appropriate review periods: allow time for difficult work while identifying persistent problems.
• Open entry routes: provide opportunities for early-career and independent researchers.
• Correction and appeal: let researchers challenge factual errors and unfair assessments.
• Responsible transitions: distinguish unsuccessful research from misconduct and avoid treating every negative result as failure.

Stable salaries also need protected research time. An employment contract dominated by other duties cannot deliver the benefits of sustained investigation merely by calling the employee a scientist.

Where AIIM fits into research funding

Science DAO’s AI Internet-Meritocracy (AIIM) explores funding based on AI-assisted assessments of documented research and software contributions. It aims to reduce reliance on traditional proposals and institutional credentials.

This creates a third distinction: funding promised future work versus rewarding contributions already made. Retrospective support could help a researcher continue working without packaging every new direction into a grant application.

AIIM is experimental. Its payments depend on available donations and eligibility; they should not be confused with guaranteed employment salaries. Its assessments are also not validated measurements of scientific or economic impact.

For a contribution-based model to provide reliable scientific support, it would need sustainable funding, trustworthy evaluation, and effective correction mechanisms. It must also address a structural limitation: people need resources to produce their first contributions, and some valuable work takes years to become visible.

AIIM is therefore a candidate for testing alongside grants and institutional employment. Its superiority in fairness, costs, or discovery output remains a hypothesis.

A practical model: support people, projects, and infrastructure

A useful funding architecture would separate three decisions:

1. Support researchers: provide sustained income and protected time for scientific work.
2. Finance projects: allocate additional resources for experiments, travel, specialist staff, and other specific needs.
3. Maintain shared infrastructure: fund facilities, repositories, datasets, and software used across projects.

For a mathematician working mainly with existing resources, protected time may be the central need. For an experimental team, it may be only one component of a much larger budget. Neither should be forced into a funding model designed around the other.

A pilot comparing funding approaches should track administrative time, continuity, substantive research quality, openness, and access for newcomers over an appropriate period. Publication totals alone would be an inadequate verdict.

The strongest case for scientific salaries is that capable researchers need continuity and room to investigate. The strongest case for research grants is that specific activities need resources and scrutiny. Science benefits when funding preserves both the person’s capacity to discover and the practical means to do the work.