In 2013, acclaimed filmmaker and author Errol Morris ran a bold experiment. With the collusion of the New York Times, he asked 45,000 readers to take an online test. The test allegedly measured whether or not readers were optimists or pessimists. But in reality, Morris was trying to find out if the typeface a statement was written in had any impact on a reader’s willingness to agree with that statement. Simply put, are some typefaces more believable than others? The answer is yes. Baskerville … was statistically more likely to influence the minds of readers.
As one might expect, the results of Morris’s experiment got a lot of attention from both inside and outside the design world. As much as I love to see typography leak into the mainstream media, I found this search for the most truthworthy typeface to be especially untrustworthy, despite coming from a man whose work I otherwise respect. What disappointed me most was the weak methodology behind the experiment. The typefaces selected for the test were Baskerville, Computer Modern, Georgia, Helvetica, Comic Sans and Trebuchet. Why these six? Morris tells Co.Design:
“I’d like to say there’s something that really went into it, that was scientific, and I wanted a combination of serif and sans serif fonts, so we picked three of each. It could have been more, but we picked six. Why those six? Your guess is as good as mine.”
Yikes. That is less than rigorous. It’s not so surprising Baskerville performed best among this bunch. Only Georgia and Helvetica were any competition at all — the others being very poorly drawn, much less readable, not intended for long texts, or all of the above.
Co.Design concludes their Morris interview with this question: “have the results of your Baskerville experiment changed the typefaces you use?”
“Oh yes, I’m drinking my own Kool-Aid now. I used to write all of my manuscripts in Bembo. Now I write them in Baskerville.”
Um, but Bembo was not even part of your test! What if Bembo is really the most trustworthy of all?!
Oh well, at least Pentagram made something handsome out of this pile of poop:
“The results of Morris’s experiment were published online in a two-part essay called Hear, All Ye People; Hearken, O Earth! and have now been put into print, as the 44th edition of the Pentagram Papers, the monograph that the design firm Pentagram sends to an exclusive list of individuals each year. Pentagram partner and long-time Morris collaborator Michael Bierut put together the typographically exquisite monograph, with with the help of designer Jessica Svendsen.”
To Morris’ credit, once you get past the triviality of the experiment, his essay tells a fascinating and humanizing tale of the man behind the 18th-century typeface, John Baskerville. I was very pleased to see that Bierut and Svendsen added some authenticity by using the only digital Baskerville that comes anywhere close to Mr. Baskerville’s original type: František Štorm’s Baskerville Original. The family has two size-specific masters: “120” for large type, “10” for small. These “optical sizes” emulate the physical design compensations of metal type, allowing headlines to be delicate and showy, while body text is open, sturdy, and readable. All other digital Baskervilles have only one size, often traced from the face of some mid-range size, making it too feeble for text yet lacking the elegance of a titling face.
I haven’t yet seen Pentagram’s book in person, but it looks like a fun piece. I hope somewhere amid Morris’ history of Baskerville they included some facts about the actual fonts they used to set the text — facts which are potentially just as interesting to those who care to read a book about the effects of typeface choice.
Karsten Luecke provides a “Pro tip. Historic words set in historic typefaces ask for historic typography. (What about some space before the semicolon?)”
Once I’ve noticed the ‘Th’ ligature, it became really importunate. Such gimmicks are better served as discretionary ligatures.
A lovely and essential paper for anyone interested in communication and persuation through art direction. Kindly contact me if it is available for purchase or when it does becomes available for purchase. Thank you.
I think the study he ran was fine. There is no perfect way to select six typefaces, though it would certainly be entertaining to watch a group of typographers try. The correct critique isn’t that the methodology was week, but that the findings are limited to the typefaces that were tested. Baskerville is the best of that particular bunch.
I find the winners and the losers to be the least interesting art of the study. I’m far more fascinated with the finding that a reader’s perception of truth — gold has an atomic number of 79 — and optimism for the future are impacted by the choice of typeface. This isn’t the only recent study that has demonstrated some aspect of this. While typographers may be nonplused by this, it is hugely surprising to everyone else.
Thanks for your response, Kevin. Given your experience I’ve been wondering what you thought of this. I guess the main problem is that the headlines for this experiment were misleading and missed the point. Instead of “Baskerville is the most trustworthy typeface” it should be “The typeface you choose can affect your reader’s perception of truth”.
The author attributes people’s credulity judgments to the typeface design but not to the content of the passage. In this study cognition (reading, understanding, and knowing) interacts with visual perception (design features of the typeface). These two variables must be enumerated and isolated for repeatability of the study, if one wants to discern the design factors contributing to the results. Without asking each participant to evaluate all of the typeface designs, there is nothing to compare, just a vote for or against a given typeface. Or is it enough to say that Baskerville is most credulous to optimists. Is there a connection between a typeface design and optimism?
People today are acclimated to reading in the Baskerville typeface, which has been used in textbooks and books for hundreds of years—the very purpose of the typeface design was for book printing (i.e. reading).
As I continue my own research about people’e emotion responses to typeface designs, I will revisit this study for a closer reading. You can find my dissertation, “Human emotion response to typographic design” (2011), papers, and presentations at www.bethekoch.com. My dissertation has an extensive review of methodology of typeface studies, and I welcome collaborators in the research.
I totally agree with Stephen Coles. Why wasn’t Bembo included in the set of 6 typefaces? Also, I really despise when Comic Sans is thrown into the mix. It’s obvious that it’s not a convincing typeface so why include it? WHY
Kevin Larson—great to see you on the comment list! Do you have any references to other studies like this—that show how typographical choices impact/influence reader perceptions (of truth)? I’d be very interested to read more along these lines of inquiry!
Contributed by Stephen Coles
Contributed by Florian Hardwig