Citation metrics have become deeply embedded in academic life. They influence promotion decisions, grant applications, university rankings, hiring processes and institutional reputation. They have also become a badge of honour for many people.
Researchers are often encouraged to increase the visibility and impact of their work, while universities carefully monitor citation performance across departments, faculties and institutions.
At first glance, this may appear reasonable. Citations can indicate impact, influence, relevance and engagement within a field. But what happens when researchers realise that citation systems can be manipulated?
Self-citation is not inherently wrong
Before discussing abuse, it is important to recognise that self-citation is often legitimate. Indeed, it is to be encouraged as it indicates that you have a track record in the field, you know about the topic and it demonstrates that peer review has recognised your contribution(s).
Researchers working within a specialised area will naturally build upon their own previous work. A long-term research programme may involve several related papers over many years, making self-citation both necessary and appropriate.
In many cases, earlier papers provide:
?? Foundational theories
?? Methods and datasets
?? Experimental protocols
?? Definitions or conceptual frameworks
?? Continuity across a research agenda
Indeed, if you did not cite your own previous work (and that of others) you would need to write much more than necessary. A single citation can, literally, save many pages of writing (re-) presenting what others have already done.
A researcher who never cites their previous work might actually appear unusual.
The issue is therefore not self-citation itself. The issue is when citation behaviour shifts from scholarly necessity towards unethical metric amplification.
Why would researchers over-cite themselves?
The uncomfortable reality is that the academic system rewards citations.
Researchers know this. Universities know this. Ranking organisations know this. Funding bodies know this.
Once citations become tied to career progression and institutional success, incentives inevitably emerge.
A researcher with higher citation counts may appear more influential. A stronger h-index may help during promotion or recruitment exercises. Highly cited academics may gain visibility, invitations to be key-note speakers, leadership positions or increased credibility within their discipline.
This creates a powerful temptation. If citations are rewarded, why would some researchers not try to increase them?
Sometimes the behaviour may begin innocently. A researcher cites several previous papers to establish expertise in a topic, and this is how it should be. Over time, however, the practice morphs into citing their own papers, which are not strictly related to the topic being discussed. In some cases, it gets to the (extreme) point where researchers are citing only their own work. As an example, see one of my recent posts (see https://www.linkedin.com/feed/update/urn:li:activity:7458322875511455745).
The incentives are obvious:
?? Increasing citation counts
?? Increasing h-index scores
?? Strengthening grant applications
?? Improving promotion prospects
?? Enhancing perceived expertise
?? Increasing visibility within search engines and databases
?? Supporting institutional performance indicators
In highly competitive environments, researchers may even feel under pressure to participate simply because they believe others are already doing so. That is where the issue stops being about individual behaviour and becomes a systemic problem.
The problem with metric-driven systems
Metrics change behaviour.
This is not unique to academia. It happens in business, politics, finance and sport. Once a number becomes important, people begin optimising for the number itself.
Academic publishing is no different. If citation counts become proxies for quality, then some researchers will inevitably focus on maximising citations rather than maximising contribution.
This does not always involve obvious fraud. In fact, the more concerning issue is subtle manipulation.
For example:
?? Excessively citing previous papers regardless of relevance
?? Adding unnecessary citations during revision
?? Citation agreements (of called cartels) between groups of researchers
?? Reviewers requesting citations to their own work
?? Journals encouraging citation stacking to increase impact metrics
Individually, these actions may appear small. Collectively, they can distort the scholarly record.
When metrics become detached from quality
The long-term risk is that citation metrics gradually lose meaning.
If citation counts can be strategically inflated, then they become less reliable as indicators of genuine influence or research quality. This creates several serious consequences.
Universities may make hiring or promotion decisions based on distorted data. Funding agencies may allocate resources using unreliable indicators. Researchers who behave ethically may feel disadvantaged compared to those who aggressively optimise metrics.
Eventually, trust in the system itself begins to weaken.
Ironically, excessive self-citation may also damage the credibility of the researcher involved. Academic communities are often smaller than they appear, and patterns of unusual citation behaviour rarely go unnoticed forever.
A short-term increase in metrics may therefore create long-term reputational risk for the individual researcher.
The AI problem that few people are discussing
There is now another dimension to this issue.
Artificial intelligence systems are increasingly trained on the academic literature itself.
This means citation manipulation does not simply affect rankings or promotion exercises. It may also affect the data used to train future AI systems.
If low-quality, heavily manipulated or strategically inflated papers become embedded within the scholarly record, those distortions may propagate into:
?? AI-assisted literature reviews
?? Research summarisation systems
?? Recommendation engines
?? Automated knowledge extraction tools
?? Future academic search systems
In other words, citation manipulation may no longer remain confined to the academic publishing ecosystem. It may become part of the infrastructure of future knowledge systems.
The governance question
Citations, and their (possible) manipulation is ultimately a governance issue.
Most universities monitor research performance closely. Many track publication outputs, citations, impact factors and h-indexes through dashboards and reporting systems.
But how many institutions actively examine whether those metrics are trustworthy?
How many universities have clear policies on excessive self-citation, citation cartels or strategic citation manipulation?
And perhaps most importantly, if researchers are increasingly rewarded for citation performance, what behaviours should institutions realistically expect to emerge?
When universities place strong emphasis on measurable indicators, they should not be surprised when some individuals learn how to optimise those indicators.
The challenge is ensuring that the system still rewards genuine scholarship rather than strategic metric management.
Final thoughts
Self-citation is not automatically unethical. In many cases, it is entirely appropriate and academically necessary.
But when citation systems become targets for optimisation, the purpose of citations begins to change.
The danger is not simply that some researchers may inflate their numbers. The greater danger is that institutions, rankings and AI systems may increasingly rely on metrics that no longer reflect genuine academic contribution.
And once trust in research metrics begins to erode, rebuilding it may prove far more difficult than increasing a citation count.
Your thoughts
What do you think?
?? Where should the line be drawn between legitimate self-citation and strategic manipulation?
?? Should universities formally monitor excessive self-citation?
?? Are citation-based metrics still reliable indicators of research quality?
About the author
Graham Kendall is Acting Vice-Chancellor of GlobalNxt University, Malaysia, and Emeritus Professor at the University of Nottingham.
He writes regularly on research integrity, governance, higher education leadership and the changing dynamics of academic publishing.
If your institution is reviewing how it evaluates research performance, publication quality or research governance, he welcomes dialogue.