Research Bites: WTF: What The Font? – Typography and ELT, part 2

In my last blog post, I looked into the claim that certain fonts can hinder or heighten the impact materials have on learning. Most people see this debate an argument about which font is more effective: a serif font or a sans-serif one. It turns out neither has any significant impact on learning one way or another – people’s perceptions and preferences are often the basis for claiming the supremacy of a single font or font type.

If the debate were framed as a Georgia vs. Arial. vs Comic Sans debate, few would argue that Comic Sans could be taken seriously as a font to be used in materials design: it is informal, ugly, unprofessional, and even childish. Count me among the detractors. So, imagine my surprise when I found a research paper in the journal Cognition that show evidence that non-standard fonts such as Comic Sans (among others) may actually have a positive effect on readability, unlike the font’s counterparts. This is in part due to an effect caused by disfluency.

Before we go further, let’s understand the term disfluency. In linguistics and ELT, disfluency means ” interruptions in the regular flow of speech, such as using uh and um, pausing silently, repeating words, or interrupting oneself to correct something said previously” (Fraundorf and Arnold, 2014). In cognitive science, disfluency takes on a different meaning altogether: “the subjective, metacognitive experience of difficulty associated with cognitive tasks” (Yauman, Oppenheimer, Vaughan, 2010, p. 2). Essentially, it is an interruption to the flow at which we process cognitive tasks. A disfluent font is a font that requires more cognitive work to decipher because it is a non-standard typeface (e.g. Comic Sans), a non-standard color (e.g. grayscale), or made more difficult to read (but still legible) in some other way. Increasing the cognitive load (making things more difficult) is already an established precept in what helps us learn, so the idea that disfluency can lead to learning makes sense.

Below, I summarize Yauman, Oppenheimer, and Vaughan’s (2010) article on disfluency. Keep in mind, to my knowledge, this work has only been done on L1 English readers. Therefore, as a final section to this blog post, I present some ideas how this may be tested formally or informally for English Language Learners.


Diemand-Yauman, C., Oppenheimer, D. M., & Vaughan, E. B. (2011). Fortune favors the bold (and Italicized): Effects of disfluency on educational outcomes. Cognition118(1), 111-115. [link]


  • Ease of understanding from both the teacher and student’s perspective may not portray true learning of information.
  • In some cases, if it is harder to learn, it is likely to leader to long-term learning.
  • Disfluency has strong theoretical support:
    • fluency is related to confidence in being able to recall information later, and therefore less focus may be had;
    • disfluency shows the learner they do not yet know the material and therefore they must pay more attention to it.
  • Because disfluency requires simple changes to fonts, if it is supported by evidence, it requires no-cost and easy to implement educational changes that can benefit many students.

The Experiments

  • Experiment 1
    • 24 18-40 years olds paid $12 each (recruited via Princeton) participated.
    • Participants had 90 seconds to learn about three aliens and their 7 features each (21 features in total)
      • This was provided on a piece of paper presented in either fluent or disfluent conditions.
        • The fluent condition font was 16-point Arial in pure black.
        • The disfluent condition fonts included 12-point Comic Sans or Bodoni MT in 60% grayscale.
      • After 15 minutes of unrelated tasks, participants were asked random questions about the aliens’ features.
    • Those in the disfluent condition answered more questions correctly on average and this was statistically significant.


  • Experiment 2
    • To see if these results would generalize to an actual classroom, a second experiment was carried out.
    • 220 high schools students participated.
    • Teachers of six different classes sent all supplementary material to the researchers – mostly worksheets and PowerPoints.
    • Materials were randomly assigned to a fluent (control) group with no changes or the disfluent group where researchers changed fonts to Haettenschweiler, Monotype Corsiva, or Comic Sans Italicized, or the paper was moved up and down during photocopying to create disfluent yet legible text.
    • Teachers taught units as usual and gave exams as usual.
    • The exam results were collected by the researchers, as well as answers to a survey about the material preferences.
    • Students in the disfluent group scored higher (statistically significant) though there were no differences found between the fonts.
    • The survey did not show any statistical differences in whether students liked the fonts.
    • Overall conclusion: “This study demonstrated that student retention of material across a wide range of subjects (science and humanities classes) and difficulty levels (regular, Honors and Advanced Placement) can be significantly improved in naturalistic settings by presenting reading material in a format that is slightly harder to read.” (p. 4).


Many of us strive to be as professional as possible. If we quibble over things such as font choice, intentionally choosing a less-than-typical font, while seemingly trivial, requires us to betray our own self-perceived senses of professionalism, style, and aesthetics. Still, if it can help our students (and ourselves), it’s worth the sacrifice, isn’t it?

Keep in mind, this research was carried out on native English-speaking students in American universities and high schools. I was not able to find any research on non-Latin based alphabets, bilinguals, or second language learners. Will disfluency work for those who already have to do extra work decoding either a new alphabet or remapping phonemic and semantic associations to a familiar one? Perhaps. Perhaps not (see part 3 for an update on these questions). Research will show. Luckily, it’s simple research we could all do with our students.


Further Links of Interest

2 thoughts on “Research Bites: WTF: What The Font? – Typography and ELT, part 2

    • Anthony Schmidt says:

      Thanks for the link! I hadn’t run across his name before. It seems like he is just advocating legible fonts, which makes sense. The font he recommends, “Cambridge” costs $199 ( The two characters he promotes, the “a” and “g” do look a little more like handwriting and I have noticed my own daughter can’t recognize a typical typed “a” and “g”, but she is only 5. However, there are so few variations of “a” and “g” in the most popular types that it should really be no effort for learners to get used to them, especially if they are reading authentic materials, which must include authentic fonts, I suppose.

      Thanks for the link. It was a good read.

Leave a Reply