Last week, we were sent this article by a colleague in the philosophy department, entitled “Fortune favors the BOLD (and the Italicized ): Effects of disfluency on educational outcomes”. It’s an interesting short paper that describes the outcome of two experiments which support earlier research claims about the educational benefits of using disfluent, i.e.“slightly difficult to read”, fonts.
The motivation for these studies stems from earlier research into fluency, the feeling of ease we associate with a particular thinking operation. Apparently, we tend to have a bias in favour of fluency, so much so that it affects our judgment, e.g. to the extent that “stocks from fluently named companies are judged to have higher value, [driving] purchasing decisions, which inflates the actual value of stocks” (Oppenheimer, 2008).
Fluency, it is suggested, might even impair the quality of our judgement and our processing, by making us feel too much at ease. Disfluency, on the other hand can make us work that little bit harder so we overcome gut instinct and choose to problem solve more carefully (with more correct results).
According to the authors, this was demonstrated by these two experiments, especially the second one, which was set in natural conditions, i.e. in a school, over a term, without neither teachers nor students being aware of what exactly was being tested. In conclusion, the authors are hopeful that this insight can lead to “improving educational practices through cognitive interventions.”
But is disfluency just a matter of keeping learners on their toes, of putting extra cognitive hurdles in their way? Would our students be more successful with their problem sets if they were asked to juggle tomatoes while they’re trying to solve them? Apparently not: “It is worth noting that it is not the difficulty, per se, that leads to improvements in learning but rather the fact that the intervention engages processes that support learning.”
They also point out point at which disfluency turns into illegibility is still to be ascertained. It’s not, as yet, an exact science. Nevertheless, and this is the encouraging part, if simple, subtle changes in font can have a positive effect on understanding and retention of learning material, then this is something well worth investigating further – not least because such subtle changes would cost almost nothing.
As our philosophy colleague jokingly put it: “The implication is the harder Moodle is to read, the better students will learn!” Not quite, of course. But almost. Expect our Moodle to look very different in the future…
PS this was also picked up yesterday by the Today programme on Radio4: “Neuroscience blogger Jonah Lehrer discusses his own gut feeling that we remember ugly fonts much more easily.” (Though he isn’t as good as the genuine article)
Ref: Oppenheimer, D. M. (2008). The secret life of fluency. Trends in Cognitive Sciences, 12(6), 237–241.
I can’t decide whether this seems intuitive or not. On the one hand, if you have to work harder to get at something, you might well expect to it stay with you longer. On the other, the letters of the word are the medium, not the message; if the concept being described is hard to grasp, one might expect it to be retained longer once it is grasped, but it seems less obvious that difficulty in deciphering the medium itself would lead to better retention of the concept being communicated.
I’m not (yet) convinced by their concept of “disfluent fonts” though. For most of the paper they talk about “fluent” and “disfluent” fonts as if disfluency were a binary state, and furthermore we’re to believe that if they say a font is disfluent, then it is. They don’t provide any justification for this as far as I can see. Who says Monotype Corsiva is harder to read than Arial Black? (Well, apart from the fact that 12pt Corsiva actually has smaller letters than 12pt Arial Black.)
Their results actually fit nicely into this binary division, in that they find no reliable differences between the different disfluent fonts. Yet it seems to be that there must be significant differences in the level of disfluency of Hattenschweiler and Monotype Corsiva on a printed page. The fact that they lump them all together into this catch-all “disfluent” category could be masking more interesting results.
Having looked back through a several of the references now, I keep finding the authors describing fonts as “easy-to-read” or “hard-to-read” without demonstrating that it is so. I have little doubt that they can justify these claims, but it’s shame they didn’t.
Steve
Thanks for your comment, Steve! You make me want to defend the article now, of course… 🙂
So here goes,
One, on your second point, artists and typographers might disagree with you, not to mention Marshall McLuhan. But the authors don’t need to rely on it. The medium (typefont) is here part of the “how you learn”, not the “what you learn”.
Second I think it’s acceptable to define disfluency as “the subjective experience of difficulty associated with cognitive operations”, as they do. That suggests to me that fluency/ disfluency aren’t binary concepts, while a majority nevertheless agrees that for them 60% grayed out fonts are less visually clear than black fonts, and that cursive characters are a bit less fluent than straight characters. Oh, and they used 16pt Arial vs 12pt grayed out Comic Sans. You don’t want a conclusive table of all fonts checked as fluent or disfluent, do you…?
Whether it is intuitive is entirely up to you and for you – courtesy of Wikipedia: “the subjective nature of intuition limits the objectivity of what to call counterintuitive because what is counter-intuitive for one may be intuitive for another”. I guess my initial feeling was that I’ve grown up in an age of DTP and then the web, where the easier something is to read the more likely I am to read it to the end and not give up half way through. I’m not likely to retain information that I haven’t read.
Kris.
Hello,
Just to clarify, I absolutely do accept their definition of disfluency and I agree that it does not suggest a binary state. That’s what makes it unacceptable, in my eyes, that they go on to use it in a binary way when referring to fonts. It is fine to say “this font is more disfluent that that font” (well it would be if you also offered some evidence or reasoning that it was so…) It is not fine to say “this font is disfluent, this font is not” – not unless you set a threshold for fluency, have some way of measuring it, and yes, provide a table of all fonts checked as fluent or disfluent.
Steve
Interesting blog thread Sonja. I think the article has all the flaws Steve points out, but the issue in terms of its practical application to the creation and delivery of online learning materials (support environment or specific instances) is even more obtuse. So much work around readability is still modelled on a ‘traditional’ view of literacy grounded in print based delivery. Reading on the screen is still measured as that, ‘reading’ where the ability to identify, decode, assimilate and reconstruct information is radically different in new media forms. I didn’t say ‘better’, no advocacy here, just different.
Of course font is already an element in design but we should remember that there is often already a high degree of optionality in web based resources., there are links (often to nowhere), options to go deeper, back, forward and the ever present ‘create a new tab and pop off to Google to look at it a different way’ option…. These generally are absent from paper based readability accounts. Some sophisticated environments (read confusing to some) are also contextually responsive, cursors changing shape to indicate alternative content, or pop-up previews of alternative visualisations. Is text for learning the same as text for navigation; is text for information retrieval the same as text for disruptive thinking? Probably not.
Perhaps a parallel question for those of us trying to use the VLE effectively is how we present material for learning as apart from the structure of the environment itself. Do we make effective use of conceptual visualisations, of auditory stimulus? Richard Mayer’s research identified four elements of a effective mental models as ‘concise, concrete, coherent, and correct’ (Richard E. Mayer, “Models for Understanding,” Review of Educational Research, vol. 49, no. 1 (1989), pp. 43–64) and I suggest that this may well apply to text too. Do we understand the ‘significance’ of the printed journal article versus the discussion thread, the effectiveness (or not!) of PowerPoint slides in print or in presentation?
Users can manipulate the CSS/XSL in most browser contexts to ‘flatten’ presentational elements if they choose so certainly it is worth seeing what different degrees of dysfluidity look like but with user options to personalise. Then of course there is the ‘cultural’ dimension to font use……………