Do 1 in 5 People Really Have Dyslexia? What Does the Research Actually Say?
You've probably seen the claim that one in five people has dyslexia. That's 20% of the population. But is this figure accurate, and where does it come from?
I regularly come across posts on social media stating that 20% of people have dyslexia. It's often presented as an established fact, but rarely with any reference to the research behind it.
As a specialist teacher assessor, I think it's important that the information we share about dyslexia is accurate and supported by evidence.
So, I decided to look into where the 20% figure comes from and whether the research really supports it.
Where does the 20% figure come from?
The figure is widely associated with the Yale Center for Dyslexia & Creativity in the United States.
Yale states that dyslexia affects 20% of the population, or one in five children, based on findings from the Connecticut Longitudinal Study. This is a substantial research project that began in 1983 with a representative group of children starting school in Connecticut, and it has followed them for several decades.
So, the figure hasn't simply been invented. It comes from a respected research centre.
Yale's explanation for the figure is that schools often fail to identify dyslexia. It contrasts the very low rates reported by schools with the much higher rate found when every child in the study was assessed.
That's an important point in its own right. Many children with dyslexia are not identified at school. However, it doesn't necessarily mean that 20% is an established prevalence rate.
Interestingly, an earlier paper from the same research project, published by Shaywitz and colleagues in 1990, identified approximately 7.8% of children in second grade and 7.5% in third grade as having a reading disability. (The paper reports boys and girls separately; these are the combined figures.)
That's considerably lower than 20%.
Part of the difference is likely to lie in how reading disability was defined. In the 1990 study, children were identified if their reading scores were well below the level predicted by their IQ. This is known as a discrepancy definition. It has since fallen out of favour, and it isn't the basis of current definitions of dyslexia.
The 20% figure appears to reflect a broader view of who is struggling with reading. However, Yale's public explanation doesn't provide enough methodological detail to show exactly how the figure was calculated, or how it relates to the earlier estimate.
It's also worth noting that the study followed a group of children. Presenting the figure as "20% of the population" extends it to adults as well.
So, it raises a reasonable question: how was the 20% figure established, and are we comparing the same things?
What does other research tell us?
The British Dyslexia Association states that approximately 10% of the population is dyslexic, with around 4% severely affected. The distinction between those two figures is itself a recognition that dyslexia varies in severity. It is a long-standing estimate, and its original source isn't easy to trace.
Professors Margaret Snowling and Charles Hulme, both well-established researchers in literacy development and reading difficulties, argued in a 2013 review that reading skills vary continuously across the population. That makes the cut-off points used to identify reading difficulties essentially arbitrary.
A large study in France shows how much this matters. Di Folco and colleagues looked at around 25,000 pupils in their first year of secondary school. Depending on the criteria and thresholds used, estimates of dyslexia ranged from around 1% to 17%. Using what the researchers considered a reasonable set of criteria, the figure was 6.6% under DSM-5 and 3.5% under ICD-11, the two main international diagnostic systems.
That's a substantial range from a single group of children.
More recently, Wagner and colleagues (2020) examined why estimates of dyslexia prevalence vary so much. Rather than simply counting poor readers, they looked for an unexpected gap between a person's reading and their listening comprehension. In other words, difficulty with reading that isn't explained by difficulty understanding language more generally.
One of their most important findings was that dyslexia cannot be identified simply by looking at low reading scores. Their research also suggested that prevalence is better understood in relation to the severity of difficulties rather than as one fixed percentage.
In other words, there isn't necessarily a clear dividing line between people who have dyslexia and those who don't.
This doesn't mean dyslexia isn't a genuine and identifiable learning difficulty. It means that decisions about where to set the threshold for identification can affect how many people are included in prevalence estimates.
Why is it so difficult to establish an accurate figure?
One of the main difficulties is that reading and spelling abilities vary across the population.
Interestingly, some of the clearest evidence for this comes from the same Yale research team. In 1992, using data from the Connecticut Longitudinal Study, they reported that reading ability formed a continuous spread, with dyslexia at the lower end rather than as a separate group. They concluded that dyslexia isn't all-or-nothing but occurs in degrees, much like high blood pressure.
They also found that identification was unstable in the early years. Only 7 of the 25 children identified as dyslexic in first grade were still identified in third grade.
If where we draw the line determines who is counted, a single fixed percentage becomes hard to defend. And that conclusion comes from the very research programme behind the 20% figure.
Some people experience very significant difficulties learning to read and spell. Others experience less severe difficulties, which may become more apparent as the demands of education or employment increase.
Researchers also use different methods to identify dyslexia.
For example, a study might identify children whose reading scores fall below a particular level. Another might compare reading with IQ, or with listening comprehension. Others use screening questionnaires or look at particular patterns of reading and processing difficulties.
These approaches can produce very different results.
Importantly, having difficulty with reading or spelling does not automatically mean someone has dyslexia.
There are several reasons why someone might experience literacy difficulties, including language difficulties, differences in educational experience or other developmental factors. These can also occur alongside dyslexia.
Equally, some people with dyslexia develop effective strategies and achieve well academically, despite continuing difficulties with aspects of reading or spelling.
This is why we need to be careful about treating screening results, low reading scores and comprehensive diagnostic assessments as though they are interchangeable.
They aren't.
What does the latest definition of dyslexia tell us?
In February 2025, two important research papers were published following the Dyslexia Delphi Study.
This research brought together experts, including academics, specialist teachers, educational psychologists and people with dyslexia, to establish greater agreement about how dyslexia should be defined and identified.
One of the important conclusions was that dyslexic difficulties exist on a continuum and can be experienced to varying degrees of severity.
The definition also identifies difficulties with reading fluency and spelling as key markers of dyslexia, across different ages and languages.
Reading fluency refers to how accurately, quickly and automatically someone can read. A person may be able to read words correctly but still find reading unusually slow or effortful.
The definition describes dyslexia as a set of processing difficulties that affect the acquisition of reading and spelling, and recognises that other developmental difficulties frequently occur alongside it. Literacy difficulties are judged in relation to age, teaching and other attainments, rather than against IQ.
For identification, the accompanying assessment paper emphasises the importance of considering different sources of evidence and weighing up possible explanations before reaching a diagnostic conclusion.
What the Delphi Study does not do is establish that 20%, 10% or any other fixed proportion of the population has dyslexia.
So, does one in five people have dyslexia?
The answer is that the available research does not allow us to confidently state that one in five people has dyslexia.
The 20% figure is associated with a respected longitudinal research programme, but the same programme reported a much lower figure using stricter criteria, and other influential research has reported substantially lower estimates too.
It is also important not to assume that the lowest estimate must be correct.
Different studies are not always measuring exactly the same thing, and the proportion of people identified can change depending on the criteria used.
This is why I would be cautious about presenting any single percentage as an established fact.
The evidence supports the conclusion that dyslexia is relatively common, but exactly how common depends partly on how it is defined and identified.
Why does this matter?
I think there is a wider issue here about how we share information.
We often see statistics repeated on social media without any explanation of where they came from, how they were calculated or what they actually mean.
Someone shares a figure, somebody else repeats it, and eventually it becomes accepted as fact.
But the number of times something has been repeated doesn't tell us whether it's accurate.
This matters particularly when we're discussing learning difficulties, because parents, teachers and other professionals rely on this information to make decisions about children's education and support.
Of course, raising awareness of dyslexia is important.
There are children and adults whose difficulties haven't been recognised, and others who have experienced years of frustration before receiving appropriate support.
But we don't need to exaggerate prevalence figures to recognise the importance of dyslexia.
We can acknowledge that dyslexia is a significant learning difficulty, that people experience it differently and that identification and appropriate support matter, without presenting uncertain statistics as established facts.
For me, evidence-based practice isn't about finding research that supports what we already believe. It's about looking carefully at the evidence, recognising its limitations and being prepared to question claims that are widely accepted.
And sometimes, the most accurate answer is simply that we don't know the exact figure.
That's not a failure of the research. It's an acknowledgement of how complex the question is.
The important thing is that children and adults experiencing persistent literacy difficulties receive appropriate support, whether or not we can agree on an exact percentage of the population.
References
1. Shaywitz, S. E., Shaywitz, B. A., Fletcher, J. M., & Escobar, M. D. (1990). Prevalence of reading disability in boys and girls: Results of the Connecticut Longitudinal Study. JAMA, 264(8), 998–1002.
2. Shaywitz, S. E., Escobar, M. D., Shaywitz, B. A., Fletcher, J. M., & Makuch, R. (1992). Evidence that dyslexia may represent the lower tail of a normal distribution of reading ability. New England Journal of Medicine, 326(3), 145–150.
3. Snowling, M. J., & Hulme, C. (2013). Children's reading impairments: From theory to practice. Japanese Psychological Research, 55(2), 186–202.
4. Di Folco, C., Guez, A., Peyre, H., & Ramus, F. (2021). Epidemiology of reading disability: A comparison of DSM-5 and ICD-11 criteria. Scientific Studies of Reading.
5. Wagner, R. K., et al. (2020). The prevalence of dyslexia: A new approach to its estimation. Journal of Learning Disabilities, 53(5), 354–365.
6. Carroll, J. M., Holden, C., Kirby, P., Thompson, P. A., & Snowling, M. J. (2025). Toward a consensus on dyslexia: Findings from a Delphi study. Journal of Child Psychology and Psychiatry, 66(7), 1065–1076.
7. Holden, C., Kirby, P., Snowling, M. J., Thompson, P. A., & Carroll, J. M. (2025). Towards a consensus for dyslexia practice: Findings of a Delphi study on assessment and identification. Dyslexia, 31(1), e1800.
8. Yale Center for Dyslexia & Creativity. Update on the Connecticut Longitudinal Study.
9. British Dyslexia Association. What is dyslexia?
Comments