When Does a Research Incentive Start Distorting the Research It Is Supposed to Encourage?

When Does Encouraging Research Start Changing It in the Wrong Direction?

Universities, funders, governments, and research organizations use incentives for understandable reasons. They may reward publications, citations, competitive grants, collaboration, open science, societal impact, patents, postgraduate supervision, or work in strategic research areas. Without incentives, activities that institutions value may receive too little attention.

And incentives often work. If publications matter for promotion, researchers have a reason to publish. If grants are rewarded, researchers have a reason to seek funding. If institutions recognize open data, researchers may devote more effort to making data reusable.

But that creates a second question: what exactly does the incentive encourage?

The intended outcome and the rewarded indicator are rarely identical. A publication requirement may be intended to encourage productive scholarship, but what is directly counted is usually an output. A citation indicator may be intended to recognize influence, but what is measured is citation activity. A funding target may be intended to strengthen research capacity, while the visible result is grant income.

Once researchers begin optimizing for the indicator rather than the underlying purpose, an incentive that was designed to support research can start reshaping it.

The Problem Begins When the Reward Becomes a Better Target Than the Research Goal

Judge Research Incentives by the Behavior They Produce

An incentive is supposed to change behavior

It would be a mistake to treat any behavioral response as evidence that an incentive has failed. Incentives exist precisely because organizations want some activities to become more likely.

If a university provides seed grants for interdisciplinary research and more researchers form interdisciplinary teams, the incentive may be functioning exactly as intended. If promotion criteria recognize rigorous data sharing and researchers devote more effort to documentation and reusable datasets, that behavioral change may also be desirable.

The relevant question is therefore not, “Did researchers respond to the incentive?” It is, “Did their response advance the underlying objective?”

The rewarded measure is often a proxy for what an institution actually wants

Many things that research systems value are difficult to measure directly. Research quality, originality, rigor, contribution, long-term scientific influence, public benefit, mentorship, and intellectual risk do not collapse neatly into a single number.

Institutions therefore use observable indicators. Publications can indicate scholarly productivity. Citations can provide information about scholarly attention or uptake. Grant success can demonstrate the ability to secure competitive resources. Patents may indicate certain forms of technological activity.

These indicators can provide useful information. Problems arise when the proxy becomes the operational goal.

The distinction matters because researchers can often increase a proxy without proportionately increasing the underlying value it is meant to represent.

Optimization becomes risky when the indicator is easier to improve than the underlying goal

Suppose an institution wants more meaningful research and introduces a minimum annual publication target. Researchers now have a concrete threshold against which their performance is evaluated.

Some may simply become more productive. That is the intended effect.

Others may respond by dividing related findings into several smaller papers, favoring questions that can be completed quickly, avoiding long-term or uncertain projects, choosing designs that produce publishable units more predictably, or allocating less time to activities that do not count toward the target. The number of publications may rise even if the underlying improvement in research contribution is uncertain.

This is why publishing can gradually become the goal rather than the research. The proxy has become sufficiently consequential that optimizing it may be more rational for the researcher than optimizing the less directly rewarded objective.

Distortion does not require misconduct

Some discussions of research incentives focus immediately on fabrication, falsification, plagiarism, or other serious breaches. Incentive systems can certainly create pressures relevant to research integrity, but distortion can occur well before misconduct.

A researcher may comply with every formal rule while choosing safer questions, avoiding replication, abandoning locally important topics, producing minimally publishable studies, or spending disproportionate effort on activities that improve measurable performance.

These decisions may be individually rational. Indeed, that is part of the problem. A system can produce undesirable collective outcomes even when researchers are responding reasonably to the rules they have been given.

Research assessment is especially important because careers respond to it

Hiring, promotion, funding, awards, and institutional evaluation tell researchers what their systems value in practice. The Coalition for Advancing Research Assessment (CoARA) argues that assessment should recognize diverse outputs, practices, and activities that contribute to research quality and impact, with assessment based primarily on qualitative judgment and supported by responsible use of quantitative indicators.

This approach reflects a broader concern in responsible research assessment: a narrow indicator can become disproportionately influential when it controls consequential decisions.

DORA similarly advocates assessment practices that move beyond inappropriate reliance on journal-based metrics and recognize research on its own merits and a wider range of contributions. The central issue is not that quantitative information is inherently useless. It is that indicators require interpretation and should not silently become definitions of quality.

Distortion becomes more likely when one measure carries many consequences

An indicator becomes particularly powerful when the same measure affects several parts of academic life. Imagine that publication count influences promotion, bonuses, workload allocation, institutional awards, research funding eligibility, and departmental evaluation. Researchers receive the same signal repeatedly: produce more countable publications.

Even if each individual policy seems modest, their combined effect can be substantial.

This helps explain why institutional pressure can change why researchers conduct research. Incentives do not operate in isolation. Researchers encounter a portfolio of expectations, and those expectations can reinforce one another.

Thresholds can create different behavior from open-ended rewards

The design of an incentive matters. A threshold such as “publish at least two papers” creates a target that can alter behavior near that boundary. A competitive ranking creates incentives to outperform others. A financial bonus adds a direct reward. A requirement tied to continued employment carries a much stronger consequence.

The same nominal indicator can therefore exert very different pressures depending on how it is used.

Context matters as well. A publication target that is easily achievable in one field may be unrealistic in another because publication practices, team sizes, research cycles, access to data, and typical project durations differ. Uniform incentives can therefore produce unequal or unintended effects across disciplines.

An incentive can distort topic choice as well as research conduct

Researchers do not merely decide how to conduct studies after choosing questions. They anticipate what will be rewarded before projects begin.

If strategic topics receive more funding, recognition, or publication opportunities, researchers may gravitate toward those areas. That can be beneficial when incentives direct attention toward neglected societal needs. It can also narrow inquiry if researchers abandon important questions because they appear less rewardable.

This is why the decision to pursue a question because an institution wants research in that area or because a topic is more likely to attract funding deserves separate consideration. Incentives can shape the research portfolio before any data are collected.

Responsible incentives should preserve room for judgment

Research is heterogeneous. A theoretical paper, longitudinal cohort, ethnographic study, randomized trial, software contribution, historical archive, systematic review, dataset, and community-engaged project may require very different timelines and produce different forms of contribution.

Assessment systems that recognize this diversity may reduce pressure to force unlike research into a single model of success. CoARA's Agreement on Reforming Research Assessment explicitly emphasizes diverse outputs, practices, and activities and recognizes that assessment approaches need to account for differences in context.

UNESCO's Recommendation on Science and Scientific Researchers provides a complementary principle: researchers should be able to exercise intellectual freedom and should be protected from undue influences on independent judgment. Incentive design therefore has implications beyond administrative efficiency. It affects the conditions under which scholarly judgment is exercised.

How a Reasonable Publication Incentive Can Produce Unintended Research Choices

Notice what has not happened. Nobody necessarily committed misconduct. Publication itself remains valuable. The university's original objective may also have been reasonable.

The distortion lies in the gap between what the policy intended to encourage and what researchers have learned to optimize. This is also why the possibility that publication targets encourage smaller but less important questions deserves attention beyond simple publication counts.

Common Misunderstandings About Research Incentives

Look for the Point Where Optimization Starts Replacing Judgment

As a researcher, you will probably encounter incentives you cannot personally redesign. You can still examine how they affect your decisions.

Ask what you would choose if the reward were absent. This is not because the incentive must be ignored. Rather, the comparison helps reveal what it is changing.

For institutions, the corresponding question is whether an incentive remains aligned with the behavior it was designed to encourage. This requires looking beyond the headline metric. If publication numbers increase, examine what kinds of studies are being undertaken, whether disciplinary differences are respected, whether research quality is being protected, and what activities researchers are abandoning to meet the target.

That broader perspective also helps distinguish legitimate rewards for doing research from reward structures that unintentionally redefine what counts as worthwhile research.

Check Whether a Research Incentive Is Starting to Distort Behavior

Questions About Incentives, Metrics, and Researcher Behavior

Do Not Judge an Incentive Only by Whether the Number Goes Up

How to Cite This Guide

This guide is intended to be read, shared, and used in research, teaching, and academic work. If you draw on its ideas, explanations, or other content, please acknowledge the source by citing the guide. Doing so gives appropriate credit and helps your readers locate the original resource.