Perverse Incentives (in Research)

A perverse incentive in research is a structural feature of a funding, evaluation, or reward system that motivates behavior contrary to the system’s own stated purpose. The concept is not specific to academia — it is a well-documented phenomenon across policy and economics (sometimes called the cobra effect after a colonial-era bounty scheme that increased, rather than reduced, the target population) — but it has a specific and well-studied shape in research administration, where careers, funding, and institutional prestige are attached to metrics that are only loose proxies for research quality.

Why research is particularly exposed to this problem

Research quality — rigor, originality, replicability, real-world value — is difficult and slow to assess directly. Institutions and funders have historically substituted easier-to-measure proxies: publication counts, journal placement (often summarized by Journal Citation Reports impact factor), citation counts (h-index and related metrics), and grant dollars secured. This substitution is efficient for evaluators, but Goodhart’s Law — “when a measure becomes a target, it ceases to be a good measure” — applies directly: once a proxy is known to determine hiring, tenure, promotion, or funding decisions, researchers face rational pressure to optimize the proxy itself, independent of whether doing so serves the underlying research goal the proxy was meant to represent.

This dynamic is commonly summarized as “publish-or-perish”: an evaluation culture where the number and placement of publications, more than their content, determines career survival.

Documented behaviors linked to metrics-driven incentive structures

• Salami-slicing — dividing a single coherent study into the smallest number of separately publishable papers, inflating publication count without producing proportionally more knowledge. See Salami slicing.
• Self-plagiarism — reusing substantial portions of one’s own previously published text, data, or results and presenting them as new work, inflating apparent output. See Self-plagiarism.
• P-hacking — exploiting flexibility in data collection, analysis choices, or reporting (researcher degrees of freedom) until a result reaches statistical significance, because significant, novel results are more publishable than null results. See P-hacking.
• HARKing (Hypothesizing After the Results are Known) — presenting a post hoc, exploratory finding as though it had been the a priori hypothesis, which makes an incidental result look more rigorous and confirmatory than it was. See HARKing.
• Citation gaming — excessive self-citation, citation cartels, or coercive citation requests aimed at inflating citation-based metrics like the h-index rather than genuinely building on prior work.

None of these behaviors require any single researcher to be acting in bad faith in isolation — that is precisely what makes perverse incentives a systemic problem rather than simply a misconduct problem. Metascience research on these dynamics generally frames them as a predictable response to incentive structure, not solely a matter of individual ethics, which is why the primary response has focused on redesigning evaluation systems rather than only policing individual behavior.

The responsible-assessment response: DORA and CoARA

Two major coordinated efforts specifically target metrics-driven perverse incentives in research evaluation:

• The San Francisco Declaration on Research Assessment (DORA), drafted in December 2012 and released in May 2013, is a widely signed declaration recommending that funders, institutions, publishers, and researchers stop using journal-based metrics such as the Journal Impact Factor as a proxy for the quality of individual articles or researchers, and instead assess research on its own merits.
• The Coalition for Advancing Research Assessment (CoARA), formally launched in December 2022 following an agreement developed with the European Commission, Science Europe, and the European University Association, commits signatory institutions to reform research assessment criteria and processes, moving away from a narrow reliance on publication and citation counts toward qualitative and diversified evaluation of contribution and impact.

Related structural fixes that remove rather than merely discourage a specific perverse incentive include Registered Reports, where peer review and in-principle acceptance occur before results are known, decoupling publication decisions from the direction or significance of the outcome, and general responsible metrics practice — using a broader, context-appropriate basket of indicators alongside expert judgment rather than a single quantitative proxy.

Why this matters for research administration

Perverse incentives are ultimately a design problem for the institutions, funders, and publishers that set evaluation criteria, not only a behavioral problem for individual researchers. Research administrators sit at exactly the points where these criteria are written into policy — tenure and promotion guidelines, grant review rubrics, institutional rankings submissions — which is why DORA and CoARA are addressed as much to institutions and funders as to researchers themselves.