Years ago I became slightly obsessed with a 2009 MIT Open Course on teaching (here). At the time it felt (and still feels) like a vein of precious metal in a confusing and dark mine with many tunnels. One of the nuggets that came out of that I remembered as the phrase “there is data showing that people will learn more if you use a slightly hard to read font on your slides”. This is called disfluency (as in, not- or less- fluent) I have no idea if I looked for that data back then – I hope I did, but at the time I was a bit enthralled with the professor’s ideas and its possible I didn’t look it up. I have always wondered about that data. Given that many published educational psychology studies aren’t models of the scientific method, its possible my nugget is a bit of fool’s gold.
There is currently on my computer desktop a folder labeled “Thinking about thinking Fort Mt 2017” which I vaguely remember making before a Memorial Day camping trip this year. It has about 6 papers in it, all containing the author Daniel M Oppenheimer (Princeton University Dept of Psychology), and one of which appears to support my nugget. Obviously I had some free time and did some literature searching and turned these up and, naturally, wanted to be reading PDFs of them by the campfire after my kids went to bed. That was not a sarcastic sentence, btw.
In an effort to encode the findings from these papers in my memory better, I want to briefly review them, starting with the disfluency paper.
Disfluency is the previously described phenomenon that people remember information slightly better when learning if it is slightly hard to understand. As in, using a crazy font or even omitting letters in words so learners have to supply them (l_ke th_s). This appears to contradict other researcher’s observations that lowering cognitive load is the way to go (i.e. – make it easier for the learners) (I blathered on about that in this post). I am not sure there is a contradiction here, because everything I have seen about cognitive load deals with complex diagrams and how to make them more digestible, not just the ability to remember strings of text. Anyhoo.
Oppenheimer’s group does two experiments, the first involving 28 paid Princeton undergraduates that I am not interested in (small N’s bedevil the educational psychology lit imho) and the second involved 228 high school students. The protocol was straightforward – teachers from a public school in Chesterfield Ohio who taught at least two sections of the same course were enrolled and sent all supplementary material (power points & handouts) to the researchers who simply made a copy of everything in different harder to read fonts (haettenschweiler, comic sans italicized or monotype corsiva – WordPress doesn’t give me those options, so I guess you won’t remember any of this). The teachers taught those courses (AP English, Honors English, Honors Physics, Regular Physics, Honors US History, and Honors Chemistry) as they normally would, but using the normal material for one section and the harder to read material for the other. Teachers each gave their normal assessments.
The results are given in Z scores, which I had to refresh myself on (Wikipedia, I love you). Basically a Z score of 0 indicates no difference from the mean; a Z score of 1 indicates 1 standard deviation above the mean, -1 indicates 1 standard deviation below the mean.
To cut to the chase, the “disfluent” students on average scored above the mean and the “normal” students scored below. How above and below looks to me to be on the order of 10%, but I honestly need to stare at it longer to be sure of that. It looks fairly significant to me.
If it were an option, I would have a snappy wrap-up sentence in a crazy font. But it’s not.
Reference – Diemand-Yauman, Oppenheimer and Vaughan Fortune favors the Bold and italicized: Effects of disfluency on educational outcomes Cognition 118 (2011) 111–115
California Southern 