TYPOGRAPHY & DYSLEXIA
Fonts that purportedly ameliorate dyslexia have recently been featured in blogs and on-line news media: Slate (Nov. 10, 2014, with correction Nov. 12), NPR (Nov. 11, 2014), USA Today (Nov 12, 2014), The Guardian (Nov. 12, 2014), and an earlier post in Scientific American (October 26, 2011).
What is missing from these news reports is scientific evidence that special dyslexia fonts are actually better for dyslexic readers than commonly used fonts.
In preparing a literature review on dyslexia and typography for a major font vendor, I surveyed more than fifty scientific papers and books about dyslexia, paying special attention to those with typographic relevance. In the scientific literature, I found no evidence that special dyslexia fonts confer statistically significant improvements in reading speed compared to standard, run-of-the-mill fonts. Some studies found that for certain subsets of reading errors, special dyslexia fonts do reduce error rates for dyslexic readers, yet for other subsets of errors, special dyslexic fonts were no better, or in some cases worse; hence, the findings on reading errors are mixed.
For typography - the sizing, spacing, and arrangement of type, but not typeface design per se - a few scientific papers found that certain variations in typography offer statistically significant benefits to dyslexic readers.
I comment on these below, after a few remarks on dyslexia and past and current research.
In modern literate societies, the incidence of dyslexia has been estimated at somewhere between 3% and 17%, depending on methodology and diagnosis. Information on dyslexia-like phenomena prior to the 19th century is generally lacking, in part because only a fraction of most populations were literate. Dyslexia didn’t become salient until a majority of the populations of European countries like England and Germany were literate.*
It appears that dyslexia became salient only after literacy became a broad social and national goal. Dyslexia was first medically diagnosed in the last quarter of the 19th century, after European nations had begun to promote literacy. The British Education Act of 1870 instituted literacy education on a national scale, and competing nations followed suit in different degrees. Previously, literacy had not been common except among higher social classes. Estimates of degrees of literacy are somewhat uncertain, but if we are content with anecdotal evidence from novels of the era, reading was not always a favorite activity.
In Jane Austen’s Pride and Prejudice (published in 1813), heroine Elizabeth Bennet, visiting the grand country house of wealthy bachelor Charles Bingley, declines an invitation to join a card game with others of the landed gentry. Instead, she reads a book, prompting one of the card-players to express his amazement that a young woman would prefer reading to card-playing.
“Do you prefer reading to cards?” said he; “that is rather singular.”
Mr. Bingley, a courteous host, thereupon offers to fetch Miss Bennet any of his books, but he apologizes for their paucity and his own lack of reading:
“And I wish my collection were larger for your benefit and my own credit; but I am an idle fellow, and though I have not many [books], I have more than I ever looked into.”
A few decades later, across the Atlantic and half-way across America, at a far lower end of the socio-economic ladder, in Mark Twain’s Huckleberry Finn (published in 1885 but set around 1845), Huck himself tells of the few months he had been forced to attend primary school in small-town Missouri:
“I had been to school most all the time and could spell and read and write just a little, and could say the multiplication table up to six times seven is thirty-five, and I don’t reckon I could ever get any further than that if I was to live forever.”
Enough fiction, now back to science.
Some of the on-line articles about a particular font for dyslexia, named appropriately “Dyslexie”, cite scholarly theses said to support the efficacy of the font. I read two of those theses (in English) and skimmed a third (in Dutch).
Renske de Leeuw (2010), in her Master’s thesis “Special Font For Dyslexia?” at University of Twente, Holland, compared Dyslexie to Arial in reading tests of dyslexic and non-dyslexic readers, and measured (a) fluency and correctness in reading Dutch text, and (b) speed of reading aloud non-words (presumably in Dutch orthography).
In comparing Dyslexie to Arial, De Leeuw’s hypotheses were that: (1) Reading speed of dyslexics when reading text in the Dyslexie font would be greater; (2) accuracy of dyslexics would be greater when reading Dyslexie; (3) speed of normal readers would be greater for Dyslexie; (4) preferential attitudes of dyslexics would be greater for the Dyslexie font.
Her findings were that: (1) there is no significant difference for dyslexics in reading speed between Dyslexie and Arial; (2) error rates are conflicting: dyslexics make fewer errors of one kind, “Substitution” errors, on the non-word test, with the Dyslexie font than with Arial, but make more of another kind of error, “Read guessing”, with the Dyslexie font than with Arial; and male readers made more errors with the Dyslexie font than with standard fonts; (3) there is no significant difference for normal readers in reading speed between Dyslexie and Arial; (4) dyslexics showed a slight preference for the Dyslexie font but contradictorily said they would not use it because it differed from the fonts normally used in school and elsewhere.
Reading speed is a standard measure in psychophysical studies of reading conducted by Gordon Legge and summarized in Legge (2007) Psychophysics of Reading in Normal and Low Vision, Lawrence Erlbaum Associates, Mahwah, NJ. Because De Leeuw found no significant differences in speed, only small differences in error rate, and contradiction in subjective font preferences by dyslexics, this study suggests that the Dyslexie font offers little or no advantage over the widely used Arial font.
A later master’s thesis from U. of Twente by Tineke Pijpker (2013) “Reading performance of dyslexics with a special font and a colored background,” compared the fonts Dyslexie and Arial again, with the added variable of yellow versus white background paper. Here is what Pijpker says in the concluding discussion.
“The goal of this study was to examine whether a yellow background and the special font Dyslexie could improve the reading performance of dyslexics.
Hypothesis 1. The results do not show support for the hypothesis that the reading speed of dyslexics will increase when the color yellow is used as background color (H1a) and when they read a text in the special font Dyslexie (H1b).
Again, there was no improvement in reading speed with the Dyslexie font compared to Arial. Pijpker also discusses reading accuracy comparing the fonts and background color. There was no significant difference between Dyslexie and Arial on white backgrounds, but there was a significant advantage for dyslexics of a lower reading level for the Dyslexie font on yellow background compared to Arial on a yellow background. There were no significant font or color advantages for higher reading level dyslexics nor for non-dyslexic readers. It is unclear why the combination of the dyslexic font printed on yellow paper was effective.
A third paper mentioned in support of the Dyslexie font is a Bachelor’s Thesis by Lianne van Someren (2013) at the University of Amsterdam. Written in Dutch, it is a “library” thesis that reviews published papers on the subject and discusses selected theories of dyslexia, but does not report results of original experiments by the author. It does review the results of De Leeuw’s thesis, described above, but there are no new laboratory findings by the author. As I am not even a rudimentary reader of Dutch, I cannot offer an appraisal of its discussion of the theoretical causes of dyslexia.
Another font purported to benefit reading by dyslexics is “Open Dyslexic”, included in a study by Luz Rello & Ricardo Baeza-Yates, “Good Fonts for Dyslexia” (2013), in ASSETS ‘13: Proceedings of the 15th International ACM SIGACCESS Conference on Computers and Accessibility.
The fonts were Arial roman and italic, Computer Modern, Courier, Garamond (Monotype), Helvetica, Myriad, Open Dyslexic roman and italic, Times Roman and italic, and Verdana. Reading speed and fixation duration were measured by eye-tracking and comprehension tests. The study found no significant differences in reading speed for the majority of the fonts, except that Arial Italic was read significantly more slowly than Arial and Helvetica romans. In reader preferences, Verdana, Helvetica, Arial, and Times Roman were ranked highest (most preferred), while Open Dyslexic Italic, Open Dyslexic, and Garamond were ranked lowest (least preferred). The other fonts were at or near median.
Rello and Baeza-Yates conclude that: “The fonts designed specifically for dyslexia, OpenDys [Open Dyslexic] and OpenDys It. [Open Dyslexic Italic] did not lead to a better or worse readability. As in [De Leeuw’s thesis], OpenDys did not lead to a faster reading.”
In Rello and Baeza-Yates, the special dyslexic font was Open Dyslexic. In De Leeuw, the special dyslexic font was Dyslexie. Neither enabled faster reading than the other fonts tested.
Although it appears that the current crop of special fonts for dyslexia offer little or no advantage, there are, however, a few other studies involving typography and dyslexia that are more promising.
Concerning typography and dyslexia, as distinct from type design and dyslexia, here are papers or presentations that I think merit further study.
Importance of Type Size.
Beth A. O’Brien, J. S. Mansfield, G. E. Legge (2005) “The Effect of Print Size on reading speed in dyslexia.” Journal of Research in Reading. Vol. 28, No. 3. pp. 332-349
This study on the effect of type size on dyslexic reading, found that dyslexic children need larger type sizes than normal readers to achieve their maximum reading speed. As background, several studies summarized by Legge (2007) found that for normal and low-vision readers, there is a “critical print size” at which readers achieve their maximum speed. Increases above the critical print size do not significantly increase reading speed, but decreases below it substantially reduce reading speed.
Critical print sizes vary by age, grade, and reading level of children, but O’Brien et al. found that in order for young dyslexic readers to achieve their maximum reading rate, they need type approximately 32% larger than what young normal readers need to achieve their maximum reading speed.
Even with the 32% larger type sizes, dyslexic readers at their fastest still read more slowly on average than normal readers at their fastest. Hence, increased type size is not a cure for dyslexia, but it does help dyslexics improve reading speed to the point that their maximum can be reached.
This is an important finding especially for e-books and e-readers on tablet computer and smart phones, where type can easily be re-sized. From a typographical point of view, a further study is desirable in order to disambiguate the effect of x-height versus type body size. That is, would choosing a font with a 32% larger x-height be equivalent to increasing the size of a given font by 32%. Possibly not, for various reasons, but it would be good to find out for sure.
Importance of Line Length.
Matthew H. Schneps, Jenny M. Thomson, Gerhard Sonnert, Marc Pomplun, Chen Chen, Amanda Heffner-Wong (2013) “Shorter Lines Facilitate Reading in Those Who Struggle.” PLoS ONE 8(8).
Schneps et al. (2013) found that texts displayed on handheld devices (iPod, iPad) were read up to 27% faster by dyslexics when the text lines were very short in terms of number of characters, on the order of 16-18 characters per line, compared to 60 - 65 characters per line as recommended in traditional print book typography. It should be emphasized that the very short lines were composed ragged-right, not justified. Justified short lines of text would likely nullify the beneficial effects and probably would retard reading speed because of the large, variable, and unpredictable word spaces that are consequences of justification of short lines.
The magnitude of the speed increase in dyslexic reading by use of short lines is highly noteworthy when compared to other studies that find improvements of much smaller magnitudes for other typographic interventions.
Assuming these findings by Schneps et al. can easily be replicated on a range of devices, then line length would be an easy and inexpensive technique to implement on e-readers, including smart phones and pad computers, and could potentially be applied to any font.
Along with line length, other typographic factors might include word spacing and line spacing, so these latter could be investigated with line length to see if any combinations give superior results.
Importance of Letter Spacing and Crowding.
Herman Bouma & C. P. Legein (1977) “Foveal and Parafoveal Recognition of Letters and Words by Dyslexics and by Average Readers.” Neuropsychologia 1977, Vol. 15, pp. 69-80
Bouma and Legein reported that a phenomenon called “crowding” limits letter recognition more strongly in dyslexic readers than in normal readers. The visual span – number of letters recognized in a fixation – is smaller in dyslexics than in normal readers, and hence reading speed is slower because speed depends on visual span. Crowding (difficulty recognizing letters) occurs in the parafovea of the retina of the eye when visual objects are too close together in relation to their distance from the center of vision in the fovea (termed “eccentricity”).
Based on the observation of Bouma and Legein, some reading researchers have tried greater spacing between letters as a method of reducing crowding and thus improving reading by dyslexics. The results have been mixed.
Donatella Spinelli et al. (2002) “Crowding Effects on Word Identification in Developmental Dyslexia”, Cortex. Volume 38, Issue 2, found that a moderate increase in letter spacing improved dyslexics’ recognition of words.
Marco Zorzi et al. (2012) “Extra-large letter spacing improves reading in dyslexia.” Proceedings of the National Academy of Sciences, vol. 109, no. 28.found that additional letter spacing improved reading in dyslexia.
However, other investigators have not agreed with these findings. Bernt Skottun and John R. Skoyles (2012) Letter: “Interletter spacing and dyslexia.” Proceedings of the National Academy of Sciences (PNAS) vol. 109 no. 44, questioned the statistical significance of the findings by Zorzi et al.
Schneps et al. (2013) (cited above) compared normally spaced to widely letter-spaced text and found little advantage to the widely spaced variant in dyslexic reading. They point out that extra spacing may actually slow reading because the text length is greater spatially but not literally, so the eye must take more jumps and travel further to fixate on the same number of letters and words in the text.
Rosen and Pelli (2012) in “Reading faster by reducing crowding”, Poster session: Vision Sciences Society Annual Meeting 2012, 11 - 16 May 2012, comment that, “Benefits due to increased spacing are canceled by increased eccentricity.” That is, increased letter spacing pushes letters farther from the fovea and thus increases crowding of text further from the center of vision.
Rosen & Pelli in a 2012 International Patent Application No. PCT/US12/61099 “Reducing Visual Crowding, Increasing Attention and Improving Visual Span”, propose a way to reduce visual crowding by gaze contingent eye tracking and text display. This ingenious method appears to be achievable in current or near-future reading devices that combine cameras and display screens, so it will be interesting to see if the approach, if implemented, benefits dyslexic readers.
All in all, my conclusions are that certain kinds of typography do offer potential benefits for dyslexic readers, especially on electronic reading devices like tablets and e-books, but that typeface design in particular has not yet been shown to provide statistically significant benefits in reading speed for dyslexics and has shown only mixed results in reading error reduction.
* Maryanne Wolf, in Proust and the Squid: The Story and Science of the Reading Brain (2007) makes the fascinating suggestion that Plato was aware of dyslexia and argued that slow students shouldn’t be forced to learn. This would push back the observation of dyslexia more than 2,300 years, to a time when the literacy rate in Athens and its surroundings was not more than 10%, and probably less. Alas, I don’t agree with Wolf’s interpretation of Plato’s remarks, found in “The Laws”, book 7, section 810. The Greek words, and translations by Greek scholars, lead me to believe that Plato was urging that students shouldn’t be forced to learn fast *handwriting*, not fast *reading*. Plato’s words ‘takhos, kallos, and apekribosthai' (“fast, beautiful, highly wrought”) refer to handwriting, I believe, not reading. Handwriting from Plato’s era has not survived, so we don’t know what it looked like, whether beautiful or forced, or not. Today, as most teachers and professors have doubtless noticed, most students are followers of Plato, perhaps without even knowing who he was. Today’s students type their texts on computers, tablets, and smart phones, using prefabricated fonts, instead of expressing their thoughts in fast, beautiful, elegant handwriting. Unless they are taking a calligraphy course.
© Bigelow & Holmes Inc. 2014