Co.Design

Are Some Fonts More Believable Than Others?

Recently, Errol Morris pulled a covert experiment on readers of The New York Times. The result? Typefaces can sway your beliefs.

Are some fonts more believable than others? A curious experiment by documentary filmmaker Errol Morris suggests as much. After polling approximately 45,000 unsuspecting readers on nytimes.com, Morris discovered that subjects were more likely to believe a statement when it was written in Baskerville than when it was written in Computer Modern, Georgia, Helvetica, Trebuchet, or Comic Sans. Baskerville: truth’s favorite typeface?

Let’s look at how Morris got here: A frequent contributor to the New York Times’s Opinionator blog, Morris encouraged readers to peruse a passage from The Beginning of Infinity, by physicist David Deutsch, on the unlikelihood that Earth will be destroyed by an asteroid. Then, he asked them to take a survey on whether they thought Deutsch’s statement was true, and how confident they felt in that conclusion. "Are You an Optimist or a Pessimist?" the post’s headline read.

But the poll was a cover—a ruse to get at the real question, how does typography influence our perception of truth? Morris tapped animator Benjamin Berman to develop a program that altered the typeface of the Deutsch passage, such that it appeared to each reader in one of the six randomly assigned typefaces mentioned above. Cornell psychology professor David Dunning helped design the test.

The results: For every 1,000 respondents, almost five more people agreed with Deutsch’s statement when it was written in Baskerville than they did when it was written in Helvetica. That might not seem terribly impressive, but Dunning assures us that this so-called Baskerville Effect is indeed statistically significant:

It’s small, but it’s about a 1% to 2% difference — 1.5% to be exact, which may seem small but to me is rather large. You are collecting these data in an uncontrolled environment (who knows, for example just how each person’s computer is rendering each font, how large the font is, is it on an iPad or iPhone, laptop or desktop), are their kids breaking furniture in the background, etc. So to see any difference is impressive. Many online marketers would kill for a 2% advantage either in more clicks or more clicks leading to sales.

What makes Baskerville so convincing? Your guess is as good as mine. Maybe the typeface has, as Morris wonders, a sort of "religious pull" that tugs at something fundamental within us. Or maybe we’re just trained to accept some typefaces as more authoritative than others; perhaps Baskerville was the favored typeface of our childhood textbooks. Whatever the answer, Morris worries about the power of type’s invisible hand:

Truth is not typeface dependent, but a typeface can subtly influence us to believe that a sentence is true. Could it swing an election? Induce us to buy a new dinette set? Change some of our most deeply held and cherished beliefs? Indeed, we may be at the mercy of typefaces in ways that we are only dimly beginning to recognize. An effect — subtle, almost indiscernible, but irrefutably there. ("Mommy, Mommy, the typeface made me do it.")

It’d be fascinating for researchers to repeat the experiment on a larger scale, enlisting all the major fonts scattered around media today. Who knows how Baskerville would compare with Verdana or Times New Roman? It’s time we get to know our fonts better. Baskerville, stentorian and soberminded Baskerville, is a grave-faced TV anchor reading the news. Comic Sans is our gossipy idiot cousin. Morris has zeroed in on something we all implicitly knew: Typefaces have personality.

Add New Comment

27 Comments

  • Ktnagel

    The experiment showed interesting but (honestly) well-known things: the way we present messages is important to the recognition. Which font the best (among a certain set) is, will most probably be a cultural matter. The winner will not be the best one in all countries, languages, or audiences. What was found here, Baskervill being the most trusted, may be related to it's more traditional design. If you think of a chinese or japanese audience, a kanji font will of course win. In a more leisure oriented society, say among students, a more casual font might win. If technical engineers are the readers, a Sans Serif font may be ahead. Figure out other examples ;)

  • DavidLauferAIGA

    Acutally, graphic designers have long known that a serif font is easier to read in running text than a sans-serif. You may wish to investigate the book "Legibility in Print" by Miles Tinker. It is old, but many of the concepts still ring true. Serif types are harder to get right in the rasterized type on the web, so not as frequently used. I do think we are overdue for a very scientific and wide scale comparison of legibility of fonts in the new reading environments. 

  • rosswilliams

    Since this was not a random sample, but people who read the New York Times site and responded to the poll, it is statistically significant only for that population. Since Baskerville is the closest of the fonts to the one used by the New York Times, it may be that association, rather than any universal preference.

    This is bad science. 

  • Gina

    Graphic designers schooled in typography could've saved Mr. Morris a bunch of time and effort. These are well-known truths to well-educated graphic designers.

  • kemlaurin

    Fantastic piece...I have also had occasion to do similar research a long time ago on best fonts for resume writing and Garamond at the time (as taste can change)  was the winning font. Again I am not sure why but there seems to be some truth in this study that some fonts are more believable than others. 

  • Josue Oquendo

    Really interesting find and article, i wonder what would be the case if it where done with different web-safe serif fonts? for example: Georgia or Times?

  • Bruce Colthart (@bccreative)

    Any typeface that draws attention to itself, that even marginally slows down reading, will color the reader's experience causing them to subconsciously evaluate the value of what they're reading.

  • Graeme Pow

    Are you sure that the difference is in the range of 1% to 2%?  If 500 in 1000 (50% of people) agree with Deutsch's statement when it was written in Helvetica, and 505 in 1000 (50.5% of people) agree with Deutsch's statement when it was written in Baskerville, then surely the difference is only 0.5%.

  • Bonniejmoheshe

    The legibility of a font , along with the article's readability, is what affects persuasion. Design basics already dictate that serif fonts in print text are the most legible, while sans serif fonts in headlines, and non-text typograghy are easier to read, and thus grab the reader.

  • Clyde Poole

    This is truely a poor article.  If the study was conducted as described, then it is just as likely that serifed fonts are more beliveable than sanserifed ones. Or that one font was more readable on a display or in print (they didn't stay which was used).  Then there is personal preference.  For example I prefer Times New Roman which looks a lot like Baskerville and truely dislike Helvitica.  Anyway as presented in this article, nothing was proved by the experiment.

  • Jerry

    And there I made one myself as well... ;) "Did you just said" instead of "did you just say"... ;)

  • Jerry

    Bonnie, did you just said "Poor spelling and grammar definately affect believability."? With a spelling error in 'definitely'?!

  • Neuromarketing

    I agree with Nainiawo.  This is likely a cognitive fluency effect.  The dense serif font slows down the reader, who then absorbs a little more of the content.  This likely explains the small lift in persuasiveness of the Baskerville. Another explanation could be the association of a font like Baskerville with authoritative news publications that tend to use similar fonts for body copy.

  • Samveticadesign

    I find this a little inadequate. I agree that Baskerville has more authority than comic sans but not Helvetica. And other variables are too big of a factor to consider these results definitive. I think all this measured was personal opinion. Now if he'd done the same study with no typeface change to measure personal opinion, then took a group who had the same opinion and did it again with a different but related article then I might consider it adequate. This just wasn't a great experiment...

  • Nainiawo

    A 2005 study found that "texts in hard to read fonts are judged to come from less intelligent authors." (Daniel M. Oppenheimer: "Consequences of Erudite Vernacular Utilized Irrespective of Necessity: Problems with Using Long Words Needlessly" http://personal.stevens.edu/~y... ). The fonts compared were Juice ITC against Times New Roman. The paper also cites earlier results to the same effect. It is mentioned in Richard Wiseman's book "59 seconds".

  • Guest

    Baskerville is so much like from a Dickens story. It reminds me of the old signs from Victorian England; rational and very common. Somewhat like what one might find adorning 'The Old Curiousity Shop', yet with typeface that wasn't curious in and of itself. That just wasn't done back then. With shops today adorned with signs in neon and in attractive color, or with music and movement we feel we are in wonderland. With traditional typeface, perhaps some more than others there is a basic idea that all is as it seems. It is not like a theme park for the soul. Haha!

  • guest

    Everyone who works in the news and/or font biz knows serifed fonts are regarded as having more substance than san-serif fonts. I suspect that's because san-serif fonts are a relatively modern invention compared to serifed fonts, which have been around since type was first set. Why Baskerville is more highly regarded than Georgia might be because of its resemblance to Times New Roman. Or maybe it subconsciously reminds them of Sherlock Holmes!