We Call It Peer Review. Sometimes It Is Just Popularity

I tried to be calm in a webinar last week. A friend called me out. Here is what I actually think about how research funding works and who it is leaving behind.

I was invited to join a panel last week on the future of scientific funding. The other panelists were sharp people. The conversation was good. And I spent the first thirty minutes trying very hard to be measured and diplomatic and calm.

This is not who I am.

My friend Gareth Dyke, who hosted the event, knows this. He and I have worked together to train more than 7,000 healthcare professionals across the MENA region in scientific writing and peer review. We followed them until they got their papers published. Q1 journals. Some in Q2. We have seen who gets through the system and who does not. So when Gareth told me at the end to stop being polite and say what I actually think, I was relieved.

This post is what I actually think.

The system is not broken by accident

Research funding today works like a popularity contest.

If a famous university name is on your application, you get the money. If it is not, you are often ignored. This is not a controversial opinion. This is what the data shows, what researchers in underfunded regions experience every day, and what nobody in the room wants to say out loud because it implicates almost everyone sitting at the table.

The metrics we use to evaluate research quality, including journal prestige, citation counts, and institutional reputation, were built by and for a specific kind of researcher. One who writes in English as a first language. One who is affiliated with a university in North America or Europe. One who already has a publication record in journals that are themselves ranked by the same system that rewards publishing in them.

This is circular. It is also the current standard.

We do not fund the best ideas. We fund the ideas that look most like the ideas we have already funded. That is not science. That is pattern matching with a grant committee in the middle.

I am one of those people

Arabic is my first language. English is not.

I have been lucky. I was surrounded by people who supported my education and my career. Not everyone in the MEA region has that. Most do not.

When a researcher in Saudi Arabia or UAE or Egypt writes a grant application in their second language, they are already at a structural disadvantage before anyone reads a single line of their methodology. Their institution name does not carry the same automatic credibility. Their citation count is lower, not because their work is weaker, but because the journals that count citations are the same journals that are harder to access, harder to publish in, and harder to build a track record with when you are starting from outside the established network.

We call this merit-based evaluation. It is not always that.

The AI problem nobody is talking about

Everyone in the webinar agreed that AI will change research funding evaluation. More data, better signals, trajectory analysis instead of track record. I agree with all of that.

But here is what I said that I want to say more clearly now.

AI will inherit whatever bias we train it on.

If the training data is fifteen years of funded grants that went to Russell Group universities and Ivy League institutions, the model will learn that those are the signals of quality. It will recommend more of the same. Faster. At scale. With the appearance of objectivity.

There is also a methodological bias that runs deeper than institutional reputation. A randomised controlled trial conducted at a large US academic medical centre is treated as the gold standard of evidence. Real-world evidence collected in Saudi Arabia, which reflects actual patients, actual clinical practice, and actual treatment outcomes in that population, is treated as secondary. Messier. Less rigorous.

But RWE done well is not less rigorous. It is differently rigorous. And for the populations it studies, it is often more relevant than a trial that excluded those patients entirely.

If AI funding models are trained to reward RCT methodology above all else, they will systematically disadvantage researchers in regions where RCTs are harder to conduct, less generalisable to local populations, and not always the right tool for the clinical question being asked.

We will have automated the bias. That is not progress.

The citation problem

One more thing I said in the webinar that I want to expand.

We measure research impact by citation counts. How many times has your paper been referenced by other papers? High citations mean high impact. High impact means better funding chances next time.

Here is the problem. People also cite bad papers.

They cite them to criticise them. They cite them to contrast them. They cite them because they are the only paper on a topic, not because they are the best one. A paper that sparked a decade of debate about why it was wrong will have a high citation count. A paper that quietly changed clinical practice in a small hospital in Riyadh, written by a team whose institution is not indexed correctly in Scopus, will have almost none.

Citation counts measure noise. Sometimes that noise correlates with quality. Often it does not.

We have built an entire funding infrastructure on a metric that measures how much people are talking about something, not whether what they are saying is good.

---

What should change

I am not against evaluation. Research funding is finite and decisions have to be made. But the criteria for those decisions need to change.

Evaluate the team’s capacity to do the work, not just their history of being funded to do work. Look at the evidence gap, not the research question, not just the methodology proposed to answer it. What matters to patients in Egypt or Bahrain is not a lesser problem because the answer will be generated using a retrospective cohort study instead of a randomised trial. The methodological choice should follow the clinical question. Not the other way around.

Build evaluation panels that include reviewers from the regions where the research will be conducted. A grant application proposing to study rare disease burden in the Gulf should be reviewed by someone who understands GCC clinical infrastructure, GCC patient populations, and GCC data realities. Not only by someone who has never worked outside a European academic medical system.

And stop treating English fluency as a proxy for scientific rigour. It is not. It is a proxy for educational privilege and geographic luck. Those are different things.

Why I am saying this here

I run a company that helps commercial and Medical Affairs teams in the MEA region generate evidence that meets international standards. Every day I see what happens when a strong idea meets a system that was not built to recognise it.

I see researchers who have done real, meaningful work with real patients and real outcomes, who cannot get their findings funded, published, or heard, because the system’s filters were calibrated somewhere else entirely.

I am not calm about this. Gareth was right to tell me to stop pretending I was.

The research funding system is not broken for everyone. It works well for the people it was designed to serve. The question is whether we are willing to redesign it to serve everyone else too.

I think we are. But not if we keep being polite about it.

---

Nouran Hamza is the founder of MARS Global and the developer of the GCC Evidence-to-Market (E2M) Framework. She was a panelist at the Profy Predicts webinar on the future of scientific funding. The training programme referenced in this post, delivered with Gareth Dyke, resulted in more than 7,000 HCPs across MENA receiving scientific writing and peer review education. The outcomes were published in a BMC-indexed journal.