In our webinar Designing for Split-Second Clarity our UK Type Director Dr. Nadine Chahine explained how devices are changing our reading behaviors and encouraged us to think about legibility in a new way. Afterwards, she answered your questions about the research and type legibility. Here’s what she had to say.
Which is more important in legibility: x-height or character width?
"X-height is definitively very important because (specifically in Latin) it gives us the optical size. So this informs us of how much space is inside the letter. For those of you who don’t know, x-height is the size of the lowercase letter X in that particular typeface. The optical size is how big the typeface appears to be. And character width is how wide every character is drawn—for example if you compare a condensed typeface versus a regular typeface.
So the answer is a combination of both, because we are interested in the area and not simply the height. When you have typefaces with a large x-height, you usually end up with a larger area for those counters, and this makes it more legible. If you take away from that area, whether by a condensed design or by a lower x-height, then you are reducing legibility. So it’s the combination of them that is important.
If you want to go for a more condensed design, then you definitively need to have a larger x-height. If you go for a typeface with a small x-height, you definitively need a design which is wide. You want to keep in mind the width and the height of those counters, like in the letters ‘c’, ‘a’, ‘b’, ‘d’, ‘p’, ‘q’, all of those shapes that are hanging between the baseline and the x-height.
I design for medicaid and medicare recipients, which can include elderly and disabled populations as well poor readers. In my research I’ve learned that pronounced ascenders and descenders are best for the elderly. But for poor readers, large x-heights are best. So this creates a perplexing dilemma when designing for both audiences. Do you have any recommendations?
Yes. I expect that the recommendation for more pronounced ascenders and descenders is coming from the need to distinguish between similar looking characters. So, for example, if you have a single-story ‘a’ like in the Futura® typeface compared to the ‘d’, then you want that ascender to be there. However, this is contrary to what a type designer will tell you and to what the research has so far shown. For the elderly, the most important thing is the size. But I think it would be an interesting thing to research—to test typefaces with more differentiation of ascenders and descenders versus one with a larger x-height versus different sizes.
When we say poor readers, we also have to ask what is poor about them. Is it the reading skill? Their reading level? Or is it that they have low vision? Those are very different things to design for. If it’s the reading skill, I would doubt that the typeface has that much of an effect. But if it is an issue with low vision, then maybe the larger x-height could be better. There has been some research regarding low vision for children and what you would need to do to accommodate that.
The most important thing in all of this is that we don’t have enough research targeting these populations. The poor reader might be disabled visually with low vision. And if you think about accessibility, there are people who have to take medication and there have been fatalities because they did not read the name of the medicine correctly. So there’s an urgent requirement for more research into labeling, into design for the elderly, into design for the medical environment. It’s sometimes shocking that we don’t have this yet. And I do hope that we raise enough questions to highlight this so that there is more interest within the design community, within the research community, and also from a government side for more support for initiatives that would support this.
When you were doing the polarity study, did you conduct tests in both day and night environments, and did those results vary at all?
Yes. It’s not part of this presentation, but basically we did the same experiment in a simulated overcast day and in a simulated nighttime scenario. And we found that there was a bigger disadvantage for white on black at nighttime, which was a bit surprising. It seems that when you put white on black it usually performs less well than black on white, and if you do that at night, it becomes even worse.
Of course, we have not made a recommendation for the automotive market to switch the conventions that they have now because there are many good reasons why within a automotive space you do want to have white text on a black background. For one, you don’t want to have a lot of light inside the car because this will be annoying while you drive. But also because when you’re looking at the road ahead of you and it’s dark, and then you look inside at a bright, shining screen in front of you, it is difficult for your pupils to readjust back to the dark road ahead. So there’s a safety issue there.
But it just means we have to do more research there. And if we are designing for wearables or other kinds of devices, where driver distraction is not an issue, then we should definitely investigate more the effect of polarity at night.
You talked about fixations in your original study and how lower fixation rates are better. Could you clarify what you meant by the fixation rates, and why or how a lower rate is better?
A fixation is when the eye stops and focuses on a group of letters until it picks up enough information to jump on to the next collection of letters, which is usually another word. Every time that the subject’s eye stops, it stops for a certain period of time. We then look at all of those fixations in that condition and make an average out of that. So for example, as subjects interacted with our interface and its list of street names, if their eye stopped 100 times, the fixation rate was the average of those 100 durations. And average fixation is definitely a good measure of legibility.
If you are reading text and your average fixation is 300 milliseconds, and then you switch the typeface and it takes you 400 milliseconds on average to read similar text, it means this new typeface is less legible because it took you longer to read a similar amount of text of similar difficulty.
What’s the effect of comprehension on average fixation duration?
If the content is difficult, then your average fixation duration is going to go up. So, for example, if you have one passage of text that discusses weather set in the Times® typeface, 12 point, regular spacing, your average fixations might be 250 milliseconds. Then, if I give you a passage about nuclear fission, suddenly the content is very difficult to understand because it’s very technical, and the average fixation duration might jump to 350 milliseconds.
In such a case, the variable that is changing between the two conditions is the difficulty of the text, and the fixation is showing that. But in the studies I work on, because we are interested in the effect of the typeface, we keep the difficulty the same so that comprehension is not an issue.
What about the effects on movement on legibility? Reading contexts are changing, and our habits are moving toward read AND do contexts instead of read THEN do. So how would you measure the effect of reader or interface movement on the legibility of text—like while you’re running, commuting, etc?
So this is very interesting. It’s actually a good topic for a new study. We haven’t tested that specifically yet, but we have tested something else somewhat related to this—the effect of the predictability of information on screen, it’s position, and how that affects legibility and time required for reading. We do have the results for that, but unfortunately I cannot give you the results yet because it is quite literally fresh out of the oven.
So I can’t predict which typefaces, sizes, spacings, polarity, etc would perform best in those contexts, but I would expect reading speed to drop compared to stationary reading. We understand how eye movement usually works. When you’re looking at a specific word, at some point your brain starts organizing where to jump next, to what word are you going to look at next. And this is informed by what kind of information is in your peripheral vision. But if you keep moving your eyes around, that information keeps changing, and I would expect there would be a cost for that—that you would need to do more regressions to re-read words more often. But as I say we haven’t tested it yet and it’s a great suggestion.
Could you recommend some resources related to the scientific findings you presented?
Absolutely. Here are the research articles for the studies I mentioned in my talk.
- Dobel, C., Diesendruck, G., & B.lte, J. (2007). How writing system and age influence spatial representations of actions: a developmental, cross-linguistic study. Psychological Science, 18(6), 487–491.
- Maass, A., & Russo, A. (2003). Directional bias in the mental representation of spatial events: nature or culture? Psychological Science, 14(4), 296–301.
- Rayner, K. (1986). Eye movements and the perceptual span in beginning and skilled readers. Journal of Experimental Child Psychology, 41(2), 211–236.
- Reber, R., & Schwarz, N. (1999). Effects of perceptual fluency on judgments of truth. Consciousness and Cognition, 8(3), 338–342.
- Spalek, T. M., & Hammad, S. (2005). The left-to-right bias in inhibition of return is due to the direction of reading. Psychological Science, 16(1), 15–18.