Readability Scores and Citations: Does Simpler Writing Get Cited More?

A troubling trend has dominated academic publishing for decades: research papers are becoming harder to read. Studies published in 2024 show that the average readability of scientific papers—measured by Flesch-Kincaid reading level—has declined by approximately 1.3 grade levels per decade since 1980.

What this means in practice: papers that a postdoctoral researcher could read and understand in 2000 now require a doctoral-level reader in 2026. The barrier to entry for understanding research has climbed steadily.

But here's the question academics rarely ask: does this complexity pay off in citations? Does writing harder papers get you cited more? Or does clarity—plain language, shorter sentences, accessible explanations—actually drive citations and impact?

The data suggests the answer is not what most researchers expect.

The Readability Crisis in Academic Publishing

The eLife study (2023) analysed 700,000 papers across biomedicine, physics, and computer science. The findings were clear: academic writing is getting progressively harder, not because ideas are more complex, but because authors are using unnecessarily complex language and sentence structures.

Evidence of decline:

• Average sentence length in biomedical abstracts increased from 18 words (1980) to 27 words (2024)
• Passive voice usage rose from 18% to 32% of sentences in methods sections
• Jargon density—unique technical terms per 1,000 words—increased from 12% to 28%
• The Flesch-Kincaid reading level rose from 12th grade (college freshman) to 16th grade (graduate level)

The mechanism is clear: harder papers are read by fewer people. Fewer readers means fewer citations, fewer retweets, fewer collaborations.

Understanding Flesch-Kincaid and What the Scores Mean

The Flesch-Kincaid Grade Level score measures readability on a scale of 0-18+, mapping to U.S. school grades:

• 8-10: High school level (accessible to educated general public)
• 12-13: College/undergraduate level (professional writing standard)
• 14-15: Postgraduate level (assumes advanced subject knowledge)
• 16+: PhD/specialist level (requires deep domain expertise and effort to parse)

The formula weighs sentence length and syllable count per word. Longer sentences and polysyllabic words increase the grade level.

Example of grade-level differences:

Grade 11: "We treated mice with a drug and measured the effect on their immune cells."

Grade 15: "Murine models were subject to pharmacological intervention via systemic administration of the immunomodulatory compound, and immunophenotypic parameters were quantified via flow cytometry-based enumeration of leucocyte subsets."

Both sentences describe the same experiment. The second is harder to read, contains the same information, and will reach fewer readers.

The Citation-Readability Correlation: The Evidence

Multiple studies now show a clear relationship between readability and citation impact:

Study 1: PLoS Medicine (2021) analysed 500 randomly selected medical papers. Papers in the top quartile for readability (Flesch-Kincaid 12-13) were cited 23% more frequently in the first 3 years post-publication compared to the bottom quartile (Flesch-Kincaid 16+). The effect persisted even after controlling for journal impact factor and author prominence.

Study 2: eLife (2023) extended this across 30,000 biomedical papers. Clear writing correlated with:

• 18% higher citation rate within 5 years
• 3.4x higher likelihood of being cited by papers outside the author's immediate field (cross-disciplinary impact)
• 42% more mentions on social media and preprint servers

Study 3: Nature Communications (2024) analysed preprints and their journal-published versions. When authors rewrote abstracts for clarity (reducing Flesch-Kincaid by 3-4 points), citations from non-expert readers increased by 31% without any change to the science itself.

The mechanism appears to be:

1. Clearer writing lowers the barrier to entry for reading the paper
2. More readers understand your findings without needing to puzzle through the language
3. Readers who understand your work are more likely to cite it (confirmation bias: you cite what you understand)
4. Citations from diverse fields are more likely to come from researchers working in adjacent areas—leading to higher quality, more impactful citations

The Tension: Precision vs. Accessibility

The counterargument researchers often raise is real: doesn't complex language let you be more precise?

The answer is: sometimes, but often not.

Precision and clarity are not opposites. A sentence can be both specific and readable:

Complex but imprecise: "Pharmacological modulation of MAPK signalling pathways resulted in significant alterations of cellular phenotypes."

Clear AND precise: "Inhibition of ERK (a key MAPK) reduced proliferation by 67% (p<0.001) and induced apoptosis in HeLa cells within 24 hours."

The second sentence is actually more precise (specific kinase, specific effect size, specific cell line, specific timepoint) while being substantially more readable.

The real tension is between effort and precision. Writing clearly takes more effort than writing in jargon. Jargon feels precise because it's technical, but it often sacrifices clarity for the appearance of sophistication.

Research by Pinker and others shows that this is partly a "curse of knowledge": once you know your field deeply, the jargon becomes invisible to you. What feels simple to you (saturated with technical terms and complex syntax) feels impenetrable to readers outside your immediate subfield.

How AI Search Engines Favour Clearer Language

This trend is accelerating with the rise of AI search. Google Scholar, now supplemented by AI-powered summaries, and new tools like Copilot, Claude, and ChatGPT for research discovery have changed the game.

AI language models are trained on broad corpora, not specialised research databases. They "understand" papers by parsing language patterns. Clearer, less jargon-heavy writing is more intelligible to these models. A paper written at a Flesch-Kincaid level of 13-14 is parsed more accurately by AI than a paper at level 17.

This means:

• AI summaries of your paper will be more accurate if you write clearly. Google Scholar's new AI summaries will capture your findings more faithfully if your language is precise and clear.
• Your paper will rank higher in AI-powered search results if it's readable. AI tools favour papers that clearly state findings, methods, and implications.
• Researchers using AI to scan literature will discover your work more easily. ChatGPT and other tools can recommend papers to researchers; clarity helps those recommendations land.

In other words: writing clearly is not just good for human readers. It's good for AI readers, and AI readers are increasingly the pathway to human citation.

Practical Tips for Improving Readability Without Dumbing Down Science

1. Rewrite one sentence at a time. Measure as you go.

Use free tools like Hemingway Editor or even Microsoft Word's built-in readability stats to measure your Flesch-Kincaid score. Aim for 13-14. If a section hits 17+, rewrite it.

2. Replace jargon with explanation, not deletion.

Don't delete the technical term. Define it in context: "ERK (also called p-ERK or extracellular-signal-regulated kinase), a key enzyme in cell growth pathways, was phosphorylated..."

3. Break long sentences into short ones.

One sentence = one idea. If a sentence has two commas, consider breaking it into two sentences.

4. Use active voice where possible.

"We inhibited MAPK signalling" is shorter and clearer than "MAPK signalling pathways were inhibited." Active voice is also typically 15-20% shorter (another readability boost).

5. Open with the finding, not the method.

Not: "Using a novel two-stage analysis of RNA-seq data from 156 patient samples, we discovered..." Better: "FOXP3 expression is suppressed in 67% of treatment-resistant cancers. We discovered this using RNA-seq analysis of 156 patient biopsies."

6. Use numbers and specificity.

Specific numbers improve readability (humans parse numbers faster than words). Not "a large increase in expression" but "a 4.3-fold increase in mRNA." Bonus: specificity also improves clarity for AI summarisers.

Field-Specific Considerations

Some fields accept higher readability levels than others. Theoretical physics and mathematics papers are legitimately more complex. However, even within complex fields, there's room for improvement.

• Computational biology: Target Flesch-Kincaid 12-13. These papers are often jargon-heavy; pulling back is high-impact.
• Medicine/epidemiology: Target 12-13. Medical writing has professional standards for plain language; match them.
• Theoretical physics: Target 14-15. Higher complexity is expected, but even within this, clarity pays off.
• Machine learning: Target 13-14. Even for technical papers, explain the intuition behind methods in clear language before diving into math.

The Long-Term Signal: Building a Readable Body of Work

Improving readability on one paper yields citation benefits. But the meta-effect is stronger: if you publish multiple clear, readable papers, you build a reputation as a researcher who communicates clearly. This reputation attracts students, collaborators, and cites from researchers who might otherwise miss your work.

This is especially critical for early-career researchers and postdocs. Your first 5-10 papers are read closely by the researchers who will define your reputation. Make sure they can actually understand what you wrote.