Towards a Typographic Prescription for Dyslexia

This story starts a few months back when a Creative Pro reader noted that using sans serif typefaces and adding an extra pixel of character spacing in his web sites helped with his dyslexia. No sooner had I read this than I spotted a report about a study published in the Proceedings of the National Academy of Sciences (PNAS) entitled, “Extra large letter spacing improves reading in dyslexia.”

The study was led by Marco Zorzi (University of Padua) and Johannes Ziegler (Aix-Marseille University) and investigated a prediction based on earlier research: that extra wide character spacing would “ameliorate reading performance for dyslexics.” Their study supports the prediction in an unequivocal way. You can find it at

Zorzi et al. proved through scientific testing what our Creative Pro reader had observed empirically: “Extra-large letter spacing helps reading, because dyslexics are abnormally affected by crowding….” Far from adding a pixel between characters, though, the Zorzi/Ziegler team created character and word spaces you could drive a truck through, as shown in Figure 1. The test subjects were between 8 and 14 years old.

Figure 1: With character spaces almost the width of normal word spaces, and word spaces nearly 1.5 ems wide, the Zorzi team’s text allowed young readers to more easily identify individual characters while still being able to clearly perceive words as coherent entities. The test text was set in black on white in Times Roman, ragged right, on A4 paper, at 14-point with 2.5 points of added space between characters. This spacing is the equivalent of increasing normal tracking by 150/1000 em. Below this is the control text set with default word-processor spacing.

Note: The text samples in the downloadable version of the PNAS report are low-resolution bitmaps that looked bad when reproduced for this article. I re-created the type for these illustrations by superimposing new type over the bitmaps to create a precisely aligned high-res duplicates.

Young dyslexics learning to read are caught in a vicious circle, the study notes. “Reading more,” it says, “is undisputedly the most efficient intervention for dyslexia,” but reading for these kids is a torture and a humiliation.

The trick is to how to get kids over this barrier and expand their reading abilities. Traditionally, a primary therapy has been to focus on the phonological skills of dyslexics, helping them to associate certain letters and letter combinations with specific sounds in their particular language. This helps, but only very slowly, and dyslexic students typically don’t get the specialized help they need.

In addition, teaching phonics to dyslexic kids has limited efficacy because even if they can, for example, discern in isolation the characters ph (“sounds like f, students”), in a crowded (for them), normal text setting, they’re likely to misinterpret the characters’ shapes and miss the associated sound. In crowded type, recognition of characters is impaired because of visual interference between the shapes of adjoining characters. While increasing the point size of overcrowded type—that is, type that’s too tightly tracked—may help normal readers, this isn’t true for dyslexics.

The phonetic angle is interesting and important because it sheds light onto cultural aspects of dyslexia. The study was done in both Italian and French, because the former has what cognition researchers call a “transparent” writing system (that is, it has “one-to-one correspondences between letters and sounds”) whereas French (like English) is more “opaque,” with many characters pronounced inconsistently or not at all and with a greater emphasis on subtly different-sounding vowel combinations. In the Zorzi team’s tests, the Italian and French kids benefited equally from the loose spacing scheme, indicating the degree to which dyslexia is a visual problem. It wasn’t that the kids couldn’t process the text they were trying to read, they just couldn’t see it straight in the first place. Voila!, or, ecco!, the first inklings of an organized typographic prescription for alleviating the challenges faces by dyslexic readers.

The causes of dyslexia are still being argued, but, the Zorzi study notes, about 5% of the school population is affected, and the condition is debilitating. According to another report cited in the study, a dyslexic child reads in one year the same number of words that a good reader does in two days.

Although not typographers themselves, the researchers intuited what type veterans all know: that horizontal spacing should be attuned to vertical spacing, so character and word space widths go hand in hand with leading values. To that end, the first extremely spaced samples they used were set on about 32 points of lead. But when they later refined their tests, they found that adding that extra leading had no effect—all reading improvements were due solely to the increased horizontal spacing.

So how great were the improvements? Overall, the students’ comprehension levels immediately doubled when using the spaced text. Impressive. Whereas measuring comprehension of a text is fairly clear-cut (thanks to the development of methodological standards), evaluating reading speed is more difficult because it varies naturally among all readers; any reference to a “norm” in this context is of dubious value. But by comparing the known performance over time of other young dyslexic readers, the researchers determined that the kids in the test saw immediate improvements in reading speed that corresponded to the progress they would be expected to make in an entire year in normal school settings.

An insight into how much of a handicap dyslexia is for kids can be glimpsed in the course of a discussion about methodologies used by the Zorzi/Ziegler team. In preparation for one comparison, they assembled groups of kids with equivalent IQs and reading skills, half with dyslexia, half without. When matched up by skills like this, the average age in the dyslexic group was 10.9 years, while the non-dyslexics averaged 7.8 years old. In other words, in normal school settings, dyslexic kids on average—5% of the student population—may already be three years behind their peers in reading ability by the time they reach 6th grade.

The findings of the Zorzi study have yet to percolate through the community of researchers and advocates working with dyslexia and dyslexics. For example, it has yet to be squared, according to Jean Hutchins of the British Dyslexia Association (BDA), with existing observations about how dyslexics read. “It is known,” Hutchins told me, “that dyslexic people make more eye movements on a line—more saccades—taking in fewer words per movement than other readers. Wider letter spacing would make even fewer words per movement. One colleague has objected to stops [periods] in acronyms—for the benefit of Text to Speech pronunciation—because they spread out the words, which is harder for him.”

So how universal these findings are has yet to be seen. “I believe that optimal spacing might widely differ across individual dyslexics,” Zorzi acknowledged. “That’s something we are going to test in the near future. We also launched an internet-based experiment using a free iPod/iPhone app, currently available for English and French.” You can check this out for yourself at

Also yet to be quantified is how varying spacing will work for adults and to what extent individual adults’ needs and experiences will differ. Fortunately, we have, in the absence of scientific study, a potentially powerful feedback mechanism in the form of software that allows dyslexic readers to customize the presentation of their text and discover empirically what works best for each of them. Although screen displays present a range of typographical problems for most average readers, these will be far outweighed for dyslexics by the power they have to, effectively, set their own type. Social networks to the rescue.

The Zorzi study begs the question of what other typographical variables are apt to influence the dyslexic reader, and the most obvious one is typeface selection. Here, though, little formal scientific research has been done, and for the most part we have only empirical observations to rely on. The best summary I’ve found is offered by the BDA, who have also created a style guide for those designing pages and setting type for dyslexic readers. Most of what the BDA recommends is based on feedback from their own readers over the years, according to Hutchins, who compiled the BDAtech Typefaces page. One this page you can find lists and samples of typefaces that have either been designed specifically for dyslexics or have been found to be helpful in easing their reading problems. Some of these are shown in Figures 2, 3, and 4. All of these samples have been set on a lightly colored background as recommended by the British Dyslexia Association.

Figure 2: Abelardo Gonzales’s Open Dyslexic is an open-source font that takes the unusual tack of beefing up the lower ends of its characters. This adds texture to the line and breaks the monotonous balance of classically constructed faces that can cloud visual perception for dyslexics.

Figure 3: Lexia Readable, designed by Keith Bates, takes a more conservative approach to letterform design, focusing on creating distinctive forms for commonly confused characters, such as b and d.

Figure 4: Rosemary Sassoon designed her eponymous typeface for children just learning to read. But many of the same features that help youngsters to quickly identify the letters of the alphabet have also been a boon to dyslexic readers both young and old.

Despite all of these typographic resources, says Hutchins, “we do not hear many views of dyslexic people, apart from those who design fonts. The Typefaces page is by far the most visited on BDAtech web, apart from the home page, yet not one person has responded to tell us his or her preferred font style as we requested.”

As with their responses to various character and word spacing schemes, it’s reasonable to assume that dyslexic people’s reactions to various typeface designs will also vary. These are still very early days, though, and in the absence of scientific insights on the subject, trial and error will be the way forward for the time being.

Interestingly, most of the faces listed on the BDA site bear a notable resemblance to a typeface—Renner—that T. L. De Vinne cited as a paragon of readability in his 1899 work, “The Practice of Typography.” Renner (seen in Figure 5) is based on the late 15th-century faces of Franz Renner and was commissioned by De Vinne for his own press “to exemplify the belief of the writer that the legibility of print does not depend so much upon an increase in the blackness or thickness of its stems as on the entire and instant visibility of every line in every character.” De Vinne was reacting to the vestiges of Modern typeface design (as seen in Bodoni, for example) that he thought was a bane upon readers’ eyes.

Figure 5: Renner bears some striking similarities to faces designed specifically for dyslexic readers, including the unbalanced bowls of characters including b (top-heavy) and d (bottom-heavy), as well as distinctive forms for others, such as the h and the open-bowled a.

In Renner you see the features that dyslexic readers apparently find attractive. There is little contrast between thick and thin strokes, not unlike in a humanist sans serif face. There are clear distinctions between similar character shapes (I and l). Also, certain distinctive forms create a varied texture along the line (v, g, h). The set of the characters is also fairly loose and noticeably less crowded than is typical in today’s typeface designs, allowing the serifs to act better in their role as identifying features rather than as crowding elements that give dyslexics so much trouble.

I suggested to Zorzi that perhaps the choice of Times Roman in itself may have influenced the results of his tests, because having been designed for newspaper use, it’s about 10% narrower than standard text faces. This would seem to exacerbate crowding problems for dyslexics. He explained that Times was a chosen because of its ubiquity, but granted that “the use of a ‘difficult’ font might enhance the benefit of larger spacing.” What the effect of using another typeface might be has yet to be seen. Century Schoolbook, for example, was specifically designed (again with the input of Mr. De Vinne) for those learning to read.

The scientific method being as rigorous as it is, testing to determine optimal typefaces for dyslexics is not as easy as it might seem. Methodologies have become extremely sophisticated, and “one major difficulty in designing a good study,” Zorzi says, “is that an objective measure of character confusability would be necessary, especially if one is willing to explain subjective preferences.” That’s a tall order. On the other hand, most dyslexics would, I think, settle for finding a typographic prescription that works, whether or not anyone understands exactly why.


Posted on: October 3, 2012

James Felici

James Felici has worked in the publishing industry—in both editorial and production—for more than 30 years. A veteran journalist and former managing editor of Publish magazine, he has set type by hand as well as on systems from IBM, Linotype, Compugraphic, CCI, and Magna. His books include The Complete Manual of Typography (Peachpit Press), The Desktop Style Guide (Bantam/ITC), How to Get Great Type Out of Your Computer (North Light), and contributions to The Macintosh Bible (Peachpit Press). He has written for numerous publications, including PC World, Macworld, and The Seybold Report, and has been a featured speaker at Seybold Seminars, Macworld Expo, and other events worldwide.

4 Comments on Towards a Typographic Prescription for Dyslexia

  1. There are many fonts which are good for reading (especially among antiqua serif fonts where a p is not a flipped q or b) and in some ways the font is old news, there are number of fonts that are not symmetric, and the fonts that were used as a reference were designed to be symmetric for aesthetic reasons. Typography is one of the areas where skills are not always appreciated when budget cuts are on the way, “anyone can do it” or “we can all read and write” are arguments… when a project is planned.

    Note also that with the dyslexia font reading was slower for all groups, it was the difference that was less. Good kerning, tracking and leading are more important, and for the research to be relevant should have included these factors.

    On the plus is that the discussion is stirred to life.
    (Now if we talk of making problems for those with dyslexia then CAPTCHA is a real pain)

  2. By far, san serif fonts are so much easier to read than serif ones. Those little tales end up running the letters together and it is so confusing! The Open Dyslexic is very easy to read, but not an elegant font. I don’t see many using that for print, unless it is for children’s books.

  3. Check out this video about a font called Dyslexie, which explains that dyslexics ‘see letters as 3D figures’ and benefit from fonts that are heavier on the bottom and with larger counters. It also recommends making the first letter of a sentence bold, along with the ending punctuation.

    The video shows what it’s like for people with dyslexia: letters rolling and flipping… Very helpful for anyone wanting more understanding of typography and improved readability.

  4. Thanks for all your comments. I hope this article does indeed help to animate the discussion of this important issue.
    Regarding the comment that the study would be more relevant if it also investigated kerning, tracking and leading, it’s important to remember the highly focussed point of the study: to investigate the theory that expanding word spaces helps dyslexic kids read better. It’s just the first step down a road that leads to increasingly nuanced understandings of how typographic factors affect both average and dyslexic readers young and old. In fact, the study was all about tracking; it’s just grossly expanded tracking. In such an environment, kerning becomes irrelevant because with word spaces this large, kerning adjustments are too subtle to notice.
    Also, as my article points out, the study did take leading into account. Its first trials used exaggerated leading as well, on the logical assumption that vertical spacing should be harmonized to horizontal spacing. As it turned out, spreading the leading made no difference in the results, and kids did just as well with “normal” leading as long as the superwide tracking remained in effect. For optimal results with more advanced readers, though, I would expect that opening up leading somewhat in widely spaced type will create a more pleasant and efficient reading experience.
    Clearly, there are tons of questions left to investigate, such as whether exaggerated tracking using seriffed faces is enough to mitigate crowding-based comprehension problems that dyslexics now address by using sans serif faces. After all, serifs are a boon to character recognition in non-dyslexics, so might they work the same for dyslexics if used in the proper spacing context? I can see how this could be. As Dr. Zorzi points out, the challenges facing a scientific study of what makes for a suitable typeface for dyslexics are formidable. Here, I suspect, practical experience will show the way, just as it has over the centuries for mainstream typeface design development.

Leave a comment

Your email address will not be published.