The accessibility world is slowly waking up to the importance of typography in making content accessible. This is of course to be welcomed. However, there is still far too much reliance on old fonts like Arial with questionable credentials.
Recently, The Telegraph newspaper reported that UK Home Office civil servants had been told not to use Times New Roman on the grounds that it is not accessible enough. They are also banned from using italics for the same (supposed) reason.
The advice on Times New Roman is broadly correct, although I would argue that Times New Roman has probably the most legible roman numerals of any font I know (and I actually mean roman numerals, I don’t mean using the letters “i”, “v” or “x” as if they were roman numerals—a common, but very bad habit!). Also, the ‘d’ and ‘b’, and ‘q’ and ‘p’ character pairs are not mirror images of each other in Times New Roman, another legibility bonus. However, the advice on italics, as I have written about elsewhere (see the case against banning italics), is poor, and is likely to make documents less accessible, not more so.
The Telegraph article notes that the UK Supreme Court back in 2021 also dropped Times New Roman in favour of Calibri, whilst the Home Office favours Arial as an alternative. There is still no such thing as a fully accessible font, but some fonts are more accessible than others. The BBC Reith font is very much at the right end of the scale, and the Supreme Court made a reasonably good choice with Calibri. However, Arial is a poor choice, by just about any measure.
The problems with Arial
As detailed in A Guide to Understanding What Makes a Typeface Accessible, by Gareth Ford Williams of The Readability Group, Arial has small apertures (the gap between the ends of the stroke in a letter “c”, for example). It also has quite small counters (the fully enclosed part of a letter, such as the top half of the letter “e”). For these reasons, in Arial the letters “o”, “e” and “c” can easily be confused by some readers. I have experience of this myself. As a baby boomer with probably average eyesight for my age, trying to distinguish these characters in Arial at normal body text sizes can be a real problem.
To illustrate the potential impact of this, consider the following passage of text randomly selected from this week’s edition of The Economist (For Britain to grow faster it needs better managers, 31/01/2023). In the first example, the letters “o”, “e” and “c” have been randomly jumbled.
“Management is a slippcry tepic that is net much studiod by coencmists. Assossing it invelves spcaking to busincssos—and as enc dismal scicntist says, cocncmists tcnd te avcid spcaking to oompanics in much tho samo way that biclcgists tond tc avcid spcaking te ohimps.”
Now compare this to the original for readability:
“Management is a slippery topic that is not much studied by economists. Assessing it involves speaking to businesses—and as one dismal scientist says, economists tend to avoid speaking to companies in much the same way that biologists tend to avoid speaking to chimps.”
It gets worse
But small apertures and relatively small counters are not the only legibility problems with Arial. Its upper case “I” is not only the same shape as its lower case “l” (L), but they are also exactly the same height as each other—this is a real problem with words such as “Illustrator”, for example, both for accessibility reasons and for proof reading purposes. Legibility also suffers in Arial because of its below average letter spacing—its letters are too tightly packed together to be optimal for legibility. It is also a member of the “Grotesque” family of fonts that tend to have relatively even character widths which are less accessible than “Humanist” fonts which tend to have varied character widths. This is particularly a problem for people with low vision or those with cognitive disabilities.
The supposed problem of italics
Finally, Arial has no real italic. It has only a pseudo-italic (roman characters tilted forwards). You may not consider the latter to be an issue if you have been persuaded that italics are bad for accessibility. However, if you have, you should reconsider. The idea that italics are bad for dyslexic people is based on the one and only published piece of research in this field ever (Rello and Baeza-Yates, 2013), which is so flawed with respect to italics that it can safely be dismissed (see The case against banning italics for much more on this). The anecdotal evidence, on the other hand, consistently shows people with dyslexia find rich text (including italics, used correctly) good for readability. Note also that the BBC Reith font, designed with accessibility top of mind, has 8 italic variants (Light Italic, Regular Italic, Bold Italic, etc).
Calibri, on the other hand, suffers from none of the problems with Arial except that the upper case “I” and the lower case “l” are the same shape as each other. However, unlike in Arial, the lower case ascenders are noticeably taller than the cap height in Calibri, which alleviates much of the problem. Note, however, that more modern fonts, such as the free Google font Noto Sans, will often outperform both.
It is very encouraging to see organisations beginning to think about the accessibility of their typography. However, Arial really isn’t the answer to anything other than the question of how not to do it. These days there are much better options available.