03 · What You Need to Know
An Indicator Does Not Automatically Become a Good Target
Indicators, Targets, and Incentives Perform Different Functions
An indicator provides information. A target specifies desired performance. An incentive attaches consequences to achieving or failing to achieve that performance.
Indicator
A measure used to describe or assess some aspect of research or institutional performance.
Target
A desired level of performance on an indicator or related outcome.
Incentive
A reward, sanction, opportunity, or consequence attached to behavior or target attainment.
These concepts are related but should not be collapsed. A university can monitor citation performance without assigning citation quotas. It can evaluate research productivity without paying researchers per paper. It can demonstrate research capacity for quality assurance without requiring identical outputs from every faculty member.
Rankings Operate Primarily at Institutional Scale
Global university rankings generally calculate indicators using institution-level data. QS Citations per Faculty, for example, combines institutional citation data with faculty numbers. THE calculates institutional research productivity, research income, citation-based research quality, reputation, and other indicators through its own normalization and scoring procedures.
These measures describe an institution's aggregate profile. They were not necessarily designed as individual faculty performance metrics.
Translating them directly downward can therefore create a level-of-analysis problem.
An Institutional Average Does Not Define What Every Individual Should Produce
Suppose a university wants to increase publications per researcher. It might be tempting to calculate the desired institutional average and assign that number to every academic.
But researchers have different roles, disciplines, career stages, workloads, research methods, team structures, and publication cycles. Some lead large collaborative programs. Others conduct long-term individual scholarship. Some supervise doctoral researchers. Others carry substantial teaching or administrative responsibilities.
An institutional aggregate can improve through many distributions of individual contribution. Requiring everyone to produce the institutional mean assumes a uniformity that the metric itself does not require.
Rankings Normalize Because Raw Research Numbers Are Not Comparable
QS and THE apply normalization procedures precisely because universities differ in size and disciplinary profile. THE subject rankings also use different indicator weights and publication thresholds across disciplines.
If a university then ignores those differences internally and gives every department the same publication or citation target, it may discard the contextual adjustment that the ranking itself considered necessary.
This produces an odd situation: the ranking uses a sophisticated normalization procedure, while the institutional response is a spreadsheet containing one number for everyone.
Accreditation Standards Can Be Mis-Translated in the Same Way
Accreditation and quality-assurance frameworks may require evidence of research activity, faculty scholarship, research capability, postgraduate supervision, research integrity, infrastructure, or mission alignment.
Institutions sometimes convert broad standards into simple numerical targets because numbers are easier to manage and report.
Yet a standard concerning research capability cannot be reduced automatically to publication count. Capability may depend on qualified researchers, laboratories, governance, supervision, research administration, data systems, partnerships, and sustainable support.
This is why the role of research in accreditation and quality assurance should be understood from the actual applicable standard rather than institutional folklore.
Targets Can Be Useful
A critique of targets should not imply that universities should operate without measurable objectives. Targets can focus attention, coordinate activity, make expectations visible, support resource planning, and reveal whether strategic initiatives are producing change.
An institution seeking to build research activity may reasonably establish goals for doctoral completions, grant applications, research participation, publications, research training, infrastructure, open research, collaboration, or other priorities.
The relevant questions are whether the target is valid, appropriately scaled, realistically supported, and aligned with the underlying objective.
Targets Change Behavior Precisely Because They Matter
A target that carries no consequences may have little effect. Once budgets, promotion, workload, bonuses, recognition, or job security depend on it, researchers have strong reasons to respond.
This is not evidence of moral failure. It is the expected function of an incentive.
The institutional design question is therefore: if researchers respond rationally to this target, what behavior will become more attractive?
A Publication Target Can Produce More Publications
If researchers receive rewards or avoid penalties by producing publications, publication activity may increase. This can be beneficial when researchers complete worthwhile studies, disseminate previously inaccessible findings, improve writing practices, or become more research active.
Evidence from systems where research indicators became consequential in evaluation has shown that institutional research production can increase substantially.
The difficult question comes next: what else changes?
Researchers May Optimize the Target
If one publication counts as one unit regardless of its contribution, researchers have incentives to generate additional publication units. If only particular indexed outputs count, researchers may shift publication venues. If citations matter, attention may move toward topics expected to attract larger scholarly audiences.
These responses can be perfectly rational while creating divergence between target performance and underlying research value.
This is the central concern explored in what happens when publishing becomes the research goal.
Targets Can Influence Research Questions Before a Study Begins
The strongest effects of research targets may occur before any data are collected.
Researchers decide whether to pursue a difficult five-year project or several smaller studies. They decide whether a local question with modest international visibility is professionally worthwhile. They decide whether replication, methodological work, or an uncertain high-risk project fits their evaluation timeline.
A target therefore shapes not merely how completed research is reported but which research enters the pipeline in the first place.
Targets Can Favor Shorter Time Horizons
Annual research targets create a particular temporal structure. Researchers need measurable outputs within the reporting period.
Longitudinal studies, community-engaged research, infrastructure development, fundamental research, and other long-cycle projects may become comparatively difficult to justify unless institutions recognize milestones and intermediate contributions.
This can produce a portfolio optimized for annual reporting rather than long-term knowledge development.
Targets Can Redistribute Institutional Resources
Once an indicator becomes strategic, resources often follow. Internal grants, faculty positions, doctoral scholarships, workload reductions, publication incentives, laboratories, and administrative support may be directed toward units expected to contribute strongly to the target.
Concentration can sometimes be sensible. Universities cannot excel equally in every field.
The concern arises when allocation is based mainly on metric visibility rather than a transparent institutional strategy that also considers mission, disciplinary diversity, educational responsibilities, emerging capability, and societal relevance.
Targets Can Create Threshold Behavior
When evaluation uses minimum requirements, people often organize behavior around the threshold.
If two publications are required, the practical objective may become obtaining two. Once the threshold is reached, additional effort may shift elsewhere. If only specified journal categories count, work outside those categories may receive less attention regardless of its contribution.
The target has therefore created a discontinuity that may have little relationship with the continuous underlying concept of research quality or contribution.
Targets Can Encourage Strategic Attribution
Institutional research metrics depend partly on how outputs are attributed. When affiliation, authorship, or institutional reporting affects performance, organizations have reasons to ensure legitimate outputs are correctly captured.
Accurate attribution is appropriate. Problems arise when pressure encourages inappropriate authorship, misleading affiliation practices, or other attempts to obtain credit disconnected from genuine contribution.
Research-integrity and publication standards remain applicable regardless of institutional performance goals.
Targets Can Affect Collaboration
Collaboration can become attractive because it provides complementary expertise, access to data, infrastructure, or broader research questions. It may also contribute to internationalization and citation visibility.
If collaboration itself becomes a target, however, researchers may have incentives to create partnerships primarily to satisfy the indicator.
The resulting collaboration may exist on paper without the intellectual integration, reciprocal contribution, trust, and shared capability that make partnerships valuable.
One Target Can Produce Different Effects Across Disciplines
| Target |
Possible intended effect |
Possible unintended effect |
| Publications per faculty member |
Increase research activity |
Favor fields with shorter publication cycles or encourage fragmentation |
| Indexed publications |
Increase internationally visible scholarship |
Reduce incentives for legitimate outputs not covered by selected databases |
| Citation target |
Increase scholarly influence |
Favor highly citable topics or encourage inappropriate citation-oriented behavior |
| International collaboration |
Expand research networks |
Encourage nominal partnerships with limited substantive contribution |
| External funding |
Increase research resources |
Shift agendas toward funder priorities or treat grant income as research quality |
| Annual research output |
Maintain continuous productivity |
Disadvantage long-cycle or high-risk research |
Targets Can Change Researcher Motivation
A researcher may originally pursue a question because it is intellectually interesting or professionally important. Once a target is attached, satisfying the target becomes another reason for acting.
If the consequences are strong enough, research may become increasingly controlled by external contingencies. The researcher may still care about the question, but the target begins influencing which project receives time and how success is defined.
This is why institutional pressure can change research motivation.
Targets Can Become Detached From Their Original Rationale
An institution may originally create a publication target because it wants stronger research. Years later, the target may remain even though research capability, institutional mission, external standards, or ranking methodologies have changed.
Targets have organizational inertia. Once embedded in performance systems, workload policies, dashboards, and promotion criteria, they can survive the problem they were designed to solve.
Responsible management therefore requires periodic review of whether a target remains necessary and valid.
Goodhart's Law Captures the Basic Risk
The idea commonly summarized as Goodhart's law is that when a measure becomes a target, it can cease to function well as a measure. The formulation emerged from economic policy but has become widely used in discussions of metrics and organizational behavior.
The logic is highly relevant to research assessment. If publication count is useful because it provides evidence of research activity, making the count consequential can change how publications are produced. The indicator then reflects both research activity and researchers' responses to the measurement system.
This does not make the measure useless. It means interpretation must account for reactivity.
Campbell's Law Makes a Related Point About High-Stakes Indicators
Donald Campbell argued that the more a quantitative social indicator is used for social decision-making, the more pressure it faces from corruption and distortion processes and the more likely it is to distort the processes it was intended to monitor.
Research metrics are not exempt from this general problem. The higher the stakes attached to a publication, citation, funding, or ranking indicator, the more important it becomes to monitor unintended responses.
Responsible Research Assessment Reduces Reliance on Single Targets
DORA, the Leiden Manifesto, the Hong Kong Principles, and the Coalition for Advancing Research Assessment all offer related arguments for broader and more responsible research assessment.
They differ in scope and emphasis, but common themes include qualitative judgment, disciplinary context, transparency, diverse contributions, responsible quantitative indicators, and recognition of research practices that conventional output metrics may overlook.
The objective is not metric-free management. It is to avoid allowing one convenient indicator to carry more evaluative meaning than it can reasonably support.
Watch Out
Before assigning an institutional ranking or accreditation target to individual researchers, check whether the metric was designed to operate at the individual level. Aggregate indicators can produce misleading or counterproductive expectations when mechanically decomposed across departments or faculty members.
A Good Target Should Survive the Counterfactual Test
Suppose the ranking changes its methodology or the accreditation framework removes the indicator. Would achieving the target still make the university's research system better?
If increasing researcher competence, improving research integrity, strengthening doctoral supervision, developing useful infrastructure, or making credible research more discoverable remains valuable, the target is connected to something durable.
If the activity loses almost all rationale once the metric disappears, the institution may be optimizing the measurement system rather than the research.
