Is Braille Dying? What the Data Actually Say

Nobody knows. That is not a dodge: it is the finding. The figure almost always used to answer this question, "only 10% of blind people read braille", comes from a US estimate published in 1996, is quoted as if it were global and current, and a systematic review published in 2022 concluded that there is currently no source in the United States that measures braille literacy at all.

This article does not argue that braille is dying, nor that it is thriving. It argues something less comfortable and more useful: the figure the whole debate is built on cannot carry the weight we put on it. Below is exactly where it comes from, why it does not work for what we use it for, and which data can be stated with a name, a year and a population.

Disclosure: we make a braille learning app

Braillito is an app for learning braille, so we have a stake in this debate. An article from us concluding "braille is more alive than ever" is exactly what you would expect, which is why you should be suspicious of it.

We cannot remove that conflict, but we can make the article checkable:

1. Every figure links to its source, its year and its population. If we cannot say which country and which year a number belongs to, we do not publish it.
2. We do not replace one weak figure with another. Several statistics that would have suited our argument are discarded below, with the reason.
3. The conclusion does not flatter us. If we were looking for sales arguments, "a 90% braille literacy crisis" would make a far better headline than "nobody measures it".

Key points

• The 10% figure that circulates worldwide appears in The Braille Literacy Crisis in America, a March 2009 report from the National Federation of the Blind. But the report does not generate it: it cites it. Its source is an American Foundation for the Blind estimate published in 1996 on a single page of a journal.
• A systematic review of 95 publications, published in 2022 in the Journal of Visual Impairment & Blindness, traced every claim about braille literacy rates since the 1970s back to two primary sources: a 1979 National Library Service report and the American Printing House for the Blind (APH) Federal Quota data, "which should not be used as a source for braille literacy rates".
• APH says so itself, in writing, in the document it publishes every year: "The number of students who read braille in the United States CANNOT be determined from this data."
• Using the same 2024 APH census, braille readers are 7.66% or 13.92% of the total, depending on the denominator. Both figures are on the same page of the same report.
• No body measures this comparably across countries. A 2018 report for the European Blind Union surveyed nine European countries and found that "None of the countries collected national level data and statistics on the number of visually impaired children who might benefit from using Braille or who did use Braille".
• In the WHO's World report on vision (2019), more than 50,000 words on vision worldwide, the word "braille" appears once, inside a list.
• What is measured is production and lending, and there braille is not disappearing: it is moving to digital. In 2024, the US library service for blind readers circulated more than 157,000 hard-copy braille books, and its patrons downloaded nearly 450,000 in electronic braille.
• The underlying debate is real and statistics do not settle it: audio versus touch reading, changing causes of blindness, inclusive schooling and a shortage of specialist teachers. We set out each without caricature.

Where exactly the 10% figure comes from

This is the main work of this article, so we will go step by step.

The report everyone cites

The source behind almost every mention is The Braille Literacy Crisis in America: Facing the Truth, Reversing the Trend, Empowering the Blind, published on March 26, 2009 by the Jernigan Institute of the National Federation of the Blind (NFB), the largest organization of blind people in the United States. Its executive summary says:

"Fewer than 10 percent of the 1.3 million people who are legally blind in the United States are Braille readers. Further, a mere 10 percent of blind children are learning it."

That sentence, translated and trimmed, has travelled around the world for seventeen years.

It is worth being clear about what that document was and was not. It was an advocacy report, not a study. Its subtitle says so, and so does its stated goal: to double the braille literacy rate by 2015. The NFB is a serious organization with a legitimate cause, and a report calling for more braille teaching is exactly what one would expect from it. The problem is not the report: it is how it has been cited since.

The figure is not theirs: they are citing it

This is the point almost nobody checks. In the body of the report, the NFB attributes the figure explicitly:

"The American Foundation for the Blind (1996) has estimated that fewer than 10 percent of people who are legally blind in the United States and fewer than 40 percent of the estimated number who are functionally blind are Braille readers. The American Printing House for the Blind estimates the Braille literacy rate among children to be around 10 percent."

There are two figures and two different sources, and the popular version has merged them into one.

Take the first. In the report's bibliography, the full reference is:

American Foundation for the Blind. "Estimated Number of Adult Braille Readers in the United States." Journal of Visual Impairment & Blindness 90, no. 3 (May/June 1996): 287.

One page. Page 287 of a 1996 journal issue. Not a study with a described method, not a survey, not a census: an estimate published on a single page, thirty years ago.

And there is a detail always lost in the quoting: the same sentence also says that fewer than 40% of people who are "functionally blind" read braille. Two different denominators in the same sentence, with results four times apart. We will come back to that, because it is the heart of the matter.

The systematic review that went to the bottom of it

In January 2022, four researchers published in the Journal of Visual Impairment & Blindness (the American Foundation for the Blind's own journal) the work that was missing: "How Many Braille Readers? Policy, Politics, and Perception", by Rebecca M. Sheffield, Frances Mary D'Andrea, Valerie Morash and Sarah Chatfield (vol. 116, no. 1, pp. 14–25).

They carried out a systematic review of the literature, going back to the early twentieth century, and isolated 95 articles and manuscripts that claimed something specific about how many people read braille. Their abstract says:

"All identified citations used as the basis for braille literacy rates since the 1970s can be traced back to two primary sources: a National Library Service report from 1979 and the American Printing House for the Blind (APH) Federal Quota data (which should not be used as a source for braille literacy rates)."

And the practical conclusion, in the implications section:

"There is no current source of data that succinctly measures braille literacy rates in the United States, and there are no data that explain whether braille literacy rates are rising or declining."

Plainly: every claim about braille literacy rates since the 1970s goes back to two primary sources (a 1979 report and data not meant for the purpose), there is no current source measuring braille literacy in the United States, and there are no data on whether it is rising or falling.

On the AFB blog, Sheffield and Chatfield summarized the finding on March 23, 2022, quoting two lines from the paper worth repeating:

"Answering a question like 'How many braille readers are there?' requires a shared understanding of what it means to be a 'reader.' As with 'visual impairment,' there is also no single definition of 'braille literacy,' or even 'literacy'" (p. 17).

"The results of this study could be considered a cautionary tale for the field of visual impairment on using old data and repeating unsupported information" (p. 21).

An independent confirmation

That the finding comes from the AFB's own journal, the organization whose 1996 estimate it takes apart, already says a lot. But there is a second confirmation, from another discipline.

In 2023, three researchers in linguistics and psychology published in Applied Psycholinguistics "A position paper on researching braille in the cognitive sciences" (Robert Englebretson, M. Cay Holbrook and Simon Fischer-Baum). Explaining why braille is missing from the science of reading, they write:

"Sheffield et al. (2022) clearly explicate that there are no comprehensive or reliable statistics about the number of braille readers in the United States (much less in the world). They demonstrate that figures claiming 'only ten percent' of blind people read braille, which are often cited in the media, are either based on secondary sources that cite other secondary sources rather than primary counts, are based on small sample sizes that are local in nature and decades old, or are based on the extrapolation of inappropriate data sources."

Three reasons, separated: secondary sources citing secondary sources; small, local, decades-old samples; and extrapolation from data that were not meant for this. It is the most precise description we found of why the figure fails, and it is signed by nobody with a stake in the field.

The APH case: the source that forbids its own use

The second half of the 2009 sentence ("APH estimates the braille literacy rate among children to be around 10 percent") has an even clearer ending, because it can be checked in a public PDF from this year.

The American Printing House for the Blind runs the Federal Quota census every year: it registers US students who meet the definition of blindness and are therefore entitled to adapted materials, and that count determines how federal money is distributed. It is an administrative census with a funding purpose.

On the cover of the funds allocation, before any table, APH prints this notice:

"NOTICE: For Press Members and Researchers/Teachers: The number of students who read braille in the United States CANNOT be determined from this data."

And in its annual report, next to the table itself:

"Statements regarding student literacy, use of appropriate learning media, and students taught in a specific medium cannot be supported using APH registration data."

In other words: the organization whose data support half of the figure says, in writing and in capitals, that its data cannot be used to say that.

Why it does not work: the proof is in the table

The census of January 2, 2024 registered 55,382 students. This is the breakdown by primary reading medium, as APH publishes it:

Source: APH, Annual Report FY2024, Federal Quota census of January 2, 2024.

Look at the last two rows. 44.92% of registered students do not read in any medium: they are pre-readers (young children, including babies from birth) or students APH classifies as "symbolic readers and non-readers", defined as those who show no potential for traditional reading in print or braille.

If the denominator is only the students who actually read in some medium (19,477 + 4,245 + 6,779 = 30,501), the result changes. And we do not need to calculate it: APH publishes that second chart on the same page.

Both figures are correct. Both come from the same census. And they differ by almost a factor of two. Add that another 4.91% of students have braille as their secondary reading medium, and the number moves again.

That is, in one table, the whole problem of this article: a percentage without an explicit denominator means nothing. And "10% of blind people" is exactly that.

Why "10% of blind people" cannot mean one thing

There are two definitions floating inside the phrase, and neither is stable.

Who counts as blind

Legal blindness is not the absence of sight. In the United States, the administrative definition is visual acuity of 20/200 or less in the better eye with the best correction, or a visual field of 20 degrees or less. The vast majority of people who meet that definition see something, sometimes quite a lot.

Spain shows it with public numbers. The ONCE (Spain's national organization of blind people) had 71,444 members on December 31, 2024, with this makeup:

Source: ONCE, Memoria de servicios sociales para personas afiliadas 2024 (in Spanish).

Eight in ten ONCE members have severe visual impairment, not total blindness. And 47.8% are over 65, against 5.4% aged 18 or under. Any percentage that uses "blind people" as its denominator is mostly measuring older people with some remaining vision (many of whom read large print or listen), not potential school-age braille readers.

This is also, plainly, where there is genuine, deep disagreement: how much remaining vision means someone "should" read braille. It is a legitimate teaching debate, not a statistical question, and whoever settles it one way or the other will get very different figures from the same data.

Who counts as a braille reader

The second definition is slipperier still. Is a braille reader someone who reads novels at 120 words per minute? Someone who identifies a medicine by its box? Someone who labels their spices? Someone who learned it as a child and has not used it in twenty years? Someone who reads only on a braille display, never touching paper?

None of the sources we reviewed uses the same definition, and several give none. It is what both Sheffield and colleagues and Englebretson and colleagues point out: without a shared definition of "reading braille", the percentage is not even comparable with itself from one year to the next.

What other countries measure

Here we expected to find the counterweight: if the US figure is weak, maybe someone measures it better elsewhere. We went looking, and the result is itself the finding.

United Kingdom

The RNIB publishes a figure: "Around seven per cent of people who are registered blind or partially sighted use braille". It is on its reading and braille research page, published on September 12, 2022.

Three things need saying about that 7%, and all three matter:

1. The denominator is not the US one. These are people registered blind or partially sighted, a wider group than legal blindness. With a larger denominator, the percentage falls. The UK 7% and the US 10% are not comparable, and placing them side by side to describe a global trend is a mistake.
2. The source is My Voice, RNIB's own survey of blind and partially sighted people in the UK. It is a real survey, with a sample and a method, not an extrapolation.
3. The same page adds, after referring to My Voice, "For full references contact…" and an email address. The exact reference (which edition of My Voice, which question, which fieldwork year) is not published next to the figure. That the most reliable organization in Europe on this subject publishes its figure without the reference beside it is, again, the same problem in miniature.

In 2024 RNIB also published its Braille 200 report, a survey of people who already read braille about how they use it. It is a valuable document (three quarters of the sample use braille several times a week, more than half daily), but it is worth understanding what it measures: it is a survey of braille readers, not a count of how many there are. Asking braille readers how often they read says nothing about prevalence, and the report does not claim otherwise.

Europe

The European answer is the most emphatic we found. In January 2018, the Danish Association of the Blind, ICEVI and Dr Sarah Woodin published Braille Teaching and Literacy: A Report for the European Blind Union and European Commission, the result of a two-year project in nine countries: Austria, Denmark, Estonia, Finland, France, Iceland, Italy, Norway and Sweden.

The first part of its findings says:

"None of the countries collected national level data and statistics on the number of visually impaired children who might benefit from using Braille or who did use Braille. Statistics were collected for other purposes, such as to determine the number claiming welfare benefits or those eligible for services. This lack of comparable data means that the numbers of visually impaired children are not known accurately."

And later, on comparability:

"The statistics and figures on the number of visually impaired readers from the country responses are not easily comparable. Countries have developed different methods for different purposes and this resulted in different data."

When the report asked directly "how many children read and write braille at school?", these were all the answers it got:

Source: EBU / European Commission, Braille Teaching and Literacy, 2018, appendix 1, question 1.1.

France, moreover, cannot keep such a register: the report notes a legal prohibition on it. And there is a nuance the document stresses that explains half the confusion in the field: "most blind children have other impairments as well. Therefore not all children were registered as being visually impaired if this was not thought of as their main impairment".

It is worth underlining who says this. The European Blind Union is an interested party in favor of braille: its strategic plan called for every child who could benefit from braille to have access to good teaching. A report of theirs had every incentive to publish an alarming figure. Instead, it published that the figure does not exist.

Spain

We found no published estimate of the percentage of blind people who read braille in Spain. The ONCE does not publish one in its annual reports (which do detail production, requests and membership); we did not find one from Spain's braille commission; and it does not appear in the Ibero-American Braille Council documents we reviewed. What Spain does publish is the kind of data in the next section: demand and production.

The World Health Organization

In 2019 the WHO published the World report on vision, the global reference on eye health: at least 2.2 billion people with some visual impairment, with estimates by region and cause.

We downloaded the full report and went through it. In its more than 50,000 words, the word "braille" appears exactly once, in a list of adaptations: "reading using Braille, screen readers, smartphone wayfinders".

That is not a criticism of the WHO: its report is about public health, not literacy. But it is worth being clear, because the 10% figure is often presented alongside WHO data, as if it came from the same place. It does not. The WHO does not measure braille.

The more honest measure: what is produced and lent

If nobody counts readers, what can be counted? Books. And books are counted, because someone pays for them.

It is an indirect measure: it counts service, not people, and one avid reader may request fifty titles while another requests none. But it has three virtues no survey has: it is collected every year on the same basis, it is published by people who gain nothing from exaggerating it, and it can be compared with itself over time.

United States: paper holds, digital triples it

The National Library Service for the Blind and Print Disabled at the Library of Congress published this figure in its January–March 2025 newsletter, for the year 2024:

"Last year, NLS circulated more than 157,000 books in hard-copy braille and patrons downloaded nearly 450,000 books in electronic braille."

157,000 hard-copy braille books lent and nearly 450,000 electronic braille downloads, in the same year and the same service (NLS newsletter, January–March 2025).

The NLS's own timeline helps read it. Web-Braille, distributing braille books online, dates from 1999. Lending refreshable braille displays was authorized in 2016 and the first were delivered in 2017; since 2024 the NLS braille eReader has been available to any patron who requests it. In 2022 Braille-on-Demand started, letting patrons request hard-copy braille books to keep. How those devices work is in refreshable braille displays.

Spain: braille's share rises, the total falls

The ONCE publishes each year the breakdown of requests for adapted books and documents. Comparing two years:

Sources: ONCE annual social services reports for 2022 and 2024 (in Spanish).

Two years are not a trend, and we say so before anyone says it for us. But the pattern is interesting and not the one the dying-braille story predicts: total requests fall, braille's share rises, and the digital library grows in titles and downloads.

Europe: paper gives way to electronic

The 2018 report for the European Blind Union describes it this way:

"Recent years show a decline in printed Braille on paper in comparison with electronic publications. Although specialist libraries for people with visual impairments continue to produce books and articles on paper, the volume is declining overall and it is greatly outnumbered by electronic publications."

And it adds a point worth keeping even though it is uncomfortable: "Many children are expressing a preference for audio materials, especially for leisure reading".

What these three sets of data say together, and only this: paper braille is losing ground to electronic braille, and electronic braille is growing. They do not say how many people read braille, or whether there are more or fewer than before. Confusing the two is the mistake this article criticizes, and we will not make it in reverse.

The real debate, without caricature

With the statistic set aside, the underlying disagreement remains, and it is serious. These are the four arguments, as their defenders make them.

Audio and screen readers

The strong argument for audio is not that it is just as good: it is that it has hugely multiplied the amount of text available, and that it arrives sooner and costs less. The NLS and ONCE figures confirm it: audio still dominates demand.

The strong counter-argument is not sentimental. Listening to a text and reading it are not the same operation. The European Blind Union's statement on access to reading, from September 2023, puts it, drawing on Englebretson and colleagues, as braille being direct experience of literacy, while auditory forms are passive access to information. Someone who only listens does not see how a word is spelled, where a paragraph ends, how a formula is written. The 2018 European report says it more prosaically: learning braille lets users understand the formal structure of language, including spacing, formatting and grammar, and for science, mathematics and foreign languages it considers braille essential compared with audio.

It is a strong argument and, in our view, a correct one. It is also an argument with less experimental evidence than one would expect, and saying so is part of being honest.

Changing causes of blindness

This point is often caricatured and is the most important for understanding denominators. Childhood blindness in the 1950s (retinopathy of prematurity from excess oxygen, infections) is not today's. The US National Eye Institute states that "cerebral/cortical visual impairment (CVI) is a leading cause of visual impairment in children" (cvi.nih.gov), a condition of neurological origin that often comes with other disabilities.

The European report confirms it from another angle, already quoted: most blind children have other impairments as well. And the APH census shows it in raw form: 24.15% of registered students are "symbolic readers and non-readers", students the classification itself describes as having no potential for traditional reading in print or braille.

What this explains is why the "blind people" denominator does not work. What it does not explain is whether braille is taught well to those who can learn it. Using changing causes to conclude that teaching is going well would be as much of a cheat as using the 10% to conclude that it is going badly.

Inclusive schooling and teachers

The 2018 report captures the tension precisely: in all nine countries there was specialist support for children in mainstream schools, but uncertainty about whether it was enough. Braille requirements in teacher training varied enormously (university courses in Norway; in Estonia, braille was not part of training for teachers of visually impaired students). And it points to a gap that fits everything above: no country had a systematic assessment of braille reading competence and speed.

The same gap again. It is not that the data say teaching is going badly: it is that the data that would say whether it is going well or badly are not collected.

For the practical side of this debate, see teaching braille to a blind child.

How to quote a braille statistic without getting it wrong

This section is the useful part, and it comes down to five questions. If a braille statistic does not answer all five, it is not ready to publish.

And a rule that follows, which we apply here: if a figure does not survive those five questions, it is left out rather than published with caveats. Putting a doubtful number after "according to some estimates" does not fix it; it launders it.

What we could not confirm

This section is as important as the rest, and we wrote it long on purpose. This article mostly says that data are missing; it would be inconsistent not to detail which data we are missing.

And these are five things we cannot claim, however well they would fit:

1. That braille literacy is falling. Sheffield and colleagues are explicit: there are no data on whether it is rising or falling. We did not find any either.
2. That it is not falling. The same gap cuts both ways, and it would be dishonest to use it only in the direction that suits us.
3. How many people read braille worldwide. There is no defensible global figure. Englebretson and colleagues go as far as saying it makes no difference whether it is five million or five thousand, because nobody knows.
4. How many read braille in the UK, or anywhere else, beyond the figures above. Nobody we consulted publishes it.
5. Whether electronic braille is replacing paper or expanding the total. The NLS figures fit both stories, and one year cannot tell them apart.

If you know of a source that settles any of these (a national survey, a statistical series, the 1979 NLS report), we would very much like to hear about it. We will correct this article and say so.

So, is braille dying?

The honest answer has three parts, and none is a headline.

First: nobody measures it, so nobody knows. Not in the United States, not in Europe, not at the WHO. The figure used to answer it is a 1996 US estimate with no published method, and the two academic reviews that examined it agree that it does not hold.

Second: what is measured does not describe a death. Hundreds of thousands of braille books are produced and lent every year, braille's share of demand in Spain rose between 2022 and 2024, and electronic braille is growing fast. What is clearly declining is paper, which is a much smaller and much more defensible claim.

Third: the underlying debate still stands, and the figures do not settle it. How much remaining vision justifies teaching braille, whether audio makes up for what it does not give, whether inclusive schooling guarantees teachers who know it. These are teaching and policy questions, and the people living them (braille readers, braille teachers, parents deciding whether their child learns it) have arguments that are not ours to summarize on their behalf.

What is ours is not to put a 1996 statistic under them and call it evidence.

Frequently asked questions

Is it true that only 10% of blind people read braille?

No current source shows it. The figure appears in a 2009 National Federation of the Blind report, which attributes it to an American Foundation for the Blind estimate published in 1996 on a single page of the Journal of Visual Impairment & Blindness. A systematic review of 95 publications, published in 2022 in that same journal, concluded that there is currently no source measuring braille literacy in the United States and no data on whether it is rising or falling. The figure is also US-only, and is quoted as if it were global.

Where exactly does that figure come from?

From the reference "American Foundation for the Blind, Estimated Number of Adult Braille Readers in the United States, Journal of Visual Impairment & Blindness 90, no. 3 (May/June 1996): 287", cited in the NFB report The Braille Literacy Crisis in America, March 2009. The second half (10% of children) comes from the American Printing House for the Blind's Federal Quota census, which APH itself says cannot be used that way.

How many people read braille worldwide?

Nobody knows. There is no defensible global estimate. The WHO, which publishes the global figures on visual impairment, does not measure braille: in the more than 50,000 words of its 2019 World report on vision, the word appears once. A 2018 report for the European Blind Union found that none of the nine European countries studied collected national data on how many children use braille.

Is audio replacing braille?

By volume of demand, audio dominates. But what is clearly declining is paper braille compared with electronic braille, which is a different thing: in 2024, the US library service for blind readers lent 157,000 hard-copy braille books and its patrons downloaded nearly 450,000 in electronic braille. In Spain, braille's share of requests for adapted material rose from 32.4% in 2022 to 34.4% in 2024.

Why is this so hard to measure?

Because of two definitions nobody has fixed. The first is who counts as blind: most people who meet the definition of legal blindness see something, and 82.3% of the members of Spain's national blindness organization have severe visual impairment, not total blindness. The second is what "reading braille" means: labeling spice jars, identifying a medicine and reading novels are very different, and no source uses the same standard. From the same 2024 US census, braille readers are 7.66% or 13.92% depending on the denominator.

Is braille still worth learning?

That is a personal or family decision, and it is not ours to make. What can be said with sources is that braille is the only form of reading and writing in which a blind person sees spelling, punctuation and the structure of a text instead of hearing it, and that for mathematics, science and foreign languages the 2018 European report considers it essential compared with audio. To see how the system works before deciding, see the braille alphabet chart.

Who invented braille, and when?

Louis Braille, a blind French teacher, settled the six-dot cell at fifteen and published the method in 1829. His story, including the part where his own school banned the system for four years, is in who was Louis Braille.

A note on how this article is written

We avoided three temptations.

The first, the alarming headline. "90% of blind people don't read braille" is an excellent headline and would have suited us perfectly; the problem is that nobody can show it.

The second, the reassuring headline. "Braille is in good health" cannot be shown either, and coming from a company that sells braille learning it would be even more suspect.

The third, and hardest, filling the gap with the next available figure. Every time we discarded a statistic, there was the temptation to put another in its place so the paragraph would not be empty. Several times the paragraph stayed empty. We prefer that.

What we did do is leave the full trail in plain sight, with links, years and populations, so that whoever writes about this next (a journalist, a parent, a teacher, a language model) does not have to repeat a 1996 figure for lack of anything else.