Do serif fonts lead to better comprehension?

You may have heard the claim that serif fonts are easier to read and lead to a higher level of comprehension among readers. I’ve heard this repeated by various people over the years, including some I have great respect for.

It sounds like a very useful fact to know, but it’s simply untrue.

For many years I’d accepted the claim that serif fonts were easier to read and often selected serif fonts on that basis.

If you google the subject, you’ll find research going back a century or more supporting the claim. However, you’ll also find research claiming the opposite, that serif fonts are easier to read.

To add to the confusion, you’ll also find research claiming there’s no difference. (Spoiler Alert – that’s the horse you want to put your money on)

What is a serif font?

Before I get into why I think this became a widely accepted truth and why it’s not true, let’s just take a moment to clarify what a serif font is. My apologies if I’m teaching granny to suck eggs.

The illustration should be self-explanatory. The little strokes on the first font are called serifs, hence serif fonts.

Where did the claim originate?

Because there are various pieces of research that have found serif fonts are better, I can’t be certain of definitive source, but I’m pretty confident.

If you pay a visit to Amazon, you’ll likely be able to find a book called “Type & Layout: Are You Communicating or Just Making Pretty Shapes”. That’s an update of a book published in 1995, called “Type & Layout: How Typography and Design can Get Your Message Across or Get in the Way”, that was based on research the author had previously published.

It was written by Colin Wheildon who I believe was (or maybe is) an Australian newspaper editor. It covers various aspects of design and is, I believe, the source of the commonly held belief that serif fonts are easier to read

I’ve never read the book, so I’m reliant on a shorter document from 1990 named “Communicating or Just Making Pretty Shapes”. I think there’s some interesting content in the document, but I believe the claims made about serif fonts are deeply flawed. You can try googling if you want to see the complete document.

Why is the claim untrue?

I’m going to take three approaches to poo pooing the claim that serif fonts are a better choice than sans-serif fonts.

1. How could this possibly be true?

The 1990 paper presents the following table.

This is the results of comprehension levels when text was presented using a serif font and a sans-serif font.

Think about those figures for a moment.

Unless your browser settings override this site’s settings, you’re reading this article set in a sans-serif font. That should mean that only 12% of you will have a good understanding of the content. Worse still, 65% will barely understand it at all.

Does that sound remotely plausible to you?

If that was the case, the world economy would be taking a massive hit because of the volume of written communications that use sans-serif fonts and the low level of comprehension that leads to.

Huge corporations would have worked this out years ago and every advert we ever encounter would use a serif font.

That’s not the case though. Sans-serif fonts are used extensively and I don’t believe for one second that they’re adversely affecting comprehension in the way Wheildon claimed.

2. It’s old news

This research was published in 1990. Can it really be relevant to us today?

Well, no and there are good reasons for me saying that.

The research involved the subjects reading an article set in 8pt Helvetica (sans-serif font) and another set in 8pt Corona (serif font).

I worked for a newspaper publisher decades ago and 8pt is about as small as text would ever go for editorial content. Occasionally in adverts we might drop the text size to 6pt for lines like “Causes cancer” or “Will make your balls fall off”, but that was as small as it was possible to go without text becoming illegible due to ink bleed.

It made sense in 1990 to use small fonts, because small fonts meant using less paper to communicate the same amount of information.

For most of us today, this no longer applies because we’re sharing information digitally. We can make fonts as large as we like without there being any great cost overhead as a result.

The best size is subject to some discussion, but commonly 16px is considered a minimum size. On most devices that will look markedly larger and clearer than 8pt text on newsprint.

This page is set to default to 19px and Medium, one of the biggest sites sharing articles, sets their copy text at 20px.

Applying research from more than 30 years ago that used small fonts on paper just doesn’t apply to the modern day reality.

3. The science was unscientific

I read the following claim somewhere, but can’t remember where, so I can’t guarantee it’s true. The research document itself does seem to support it though.

The claim I read was that the format of the study was flawed. If correct, every single subject went through the same process.

  1. The subject was given a sheet of paper with an article printed on it in 8pt Helvetica that they read
  2. After finishing reading it, and without warning, they were asked questions about the article to assess their comprehension
  3. The subject was then given a sheet of paper with an article printed on it in 8pt Corona to read
  4. After finishing reading it, they were asked questions about the article to assess their comprehension

You can see how when reading the sans-serif text, they had no warning they would be asked questions about the article. When given the serif text, the subjects can now anticipate they will be asked questions and so were likely to have read the serif text more carefully.

That would help explain how the research came to reach what appears to be such an implausible result.

As I said, I don’t know for sure that this is how the research played out, but the document does itself offer supporting evidence for the claim.

Beyond simply testing comprehension levels, it states that participants were asked to comment about the experience of reading the sans-serif font. It then states that immediately after they were asked to read and comment on the experience of reading the serif font.

A more scientific experiment would have randomized the order that the fonts were presented, but it seems that wasn’t the case and that is likely the cause of the findings.

Should we stop using serif fonts?

When comparing serif and sans-serif fonts that have been designed for presenting body copy, we’re unlikely to ever encounter any great difference in reading ease and comprehension levels.

With older lower density displays, it was recommended to avoid serif fonts on screens, but with larger text sizes common nowadays and high pixel density screens that no longer applies.

Now, it’s more important to consider picking fonts that feel more in keeping with the subject matter.

For example, if you’re producing a sales page about an investment product aimed at an older audience, a serif font might feel a more natural choice, much like the body copy on the Wall Street Journal site.

If it’s a page selling a cutting edge device to a younger audience, a sans-serif font is going to feel like a more natural choice.

Whichever you choose, the end levels of comprehension will be largely the same if not identical.