8 May 2018
My Facebook feed has filled with people posting about this Washington Post article about a study that purportedly shows that “science” has shown that typing two spaces after a period is superior to typing just one. The number of spaces that should follow a period is one of those eternal topics of debate, with peevers and pedants on both sides assuredly proclaiming that their position is the correct one, but almost never with any evidence to show that they are, in fact, correct. So the idea that a study has definitively settled the question would be a welcome relief. The trouble is, the study in question does no such thing.
The study in question is “Are Two Spaces Better than One? The Effect of Spacing Following Periods and Commas During Reading” published last month in the journal Attention, Perception, and Psychophysics by Rebecca L. Johnson, Becky Bui, and Lindsay L. Schmitt. Unfortunately, the journal, like most academic publications, is behind a paywall, so most people don’t have access to it unless they want to fork over forty dollars. (One of the benefits of working for a major research university is that I have access to it through Texas A&M’s library.) The general public has to rely on what reporters say about it.
Now, when there is a mistake of this nature, the problem is most the fault of either the reporter or the university public relations department press release trying to make the study’s conclusion sound sexier than it really is. But in this case, the problem is with the researchers, whose conclusion does not follow from the data they present, and indeed, their experimental design precludes this study from making a definitive contribution to the question of whether or not two spaces are better than one. Blame should also go to the peer reviewers, because this paper should never have been published as it is currently written. Which is a shame, because the study does have other surprising and interesting things to say, even if it doesn’t answer the question that everyone seems to care about.
The experiment was conducted in two phases. In the first phase the participants were asked to type a short paragraph of five sentences (97 words). From the results, the participants were classified as either “one-spacers” or “two-spacers,” depending on the number of spaces they put at the end of each sentence. (All the participants used only one space following commas.) The second phase was an eye-tracking study. The participants were asked to silently read one practice and twenty test paragraphs of 71–166 words each. Each of the test paragraphs fell into one of four categories:
one space after both periods and commas;
one space after periods and two spaces after commas;
two spaces after periods and one space after commas;
two spaces after both periods and commas.
After reading each paragraph, the participants were tested on their reading comprehension and the researchers collected and analyzed data on reading speed and comprehension.
Now here is the important part. The paragraphs were presented in 14-point Courier New font with quadruple spacing between lines. Courier New is a monospaced font, like that on a typewriter, where each letter takes up the same amount of horizontal space, e.g, an < i > takes up as much space as an < m >. Most word processor fonts, and all fonts used by professional publishers, are proportionally spaced, where the horizontal space used by each letter varies with the size of the letter, e.g., an < i > takes up less space than an < m >. The published article does not specify if the paragraphs were left-aligned or justified, how long each line was, the color of the type and background, nor the resolution the monitors were set at. All of these factors can affect readability, perhaps more dramatically that post-period spacing. So, the description of the experimental design is inadequate for replication.
These choices were made for the sake of the eye-tracking software, which works best with monospaced font and quadruple spacing, despite the fact that almost no document anywhere is actually written in monospaced font with quadruple spacing. At the core of the argument of those who advocate for single spacing following periods is the claim, unsubstantiated by any good evidence but almost universally accepted, is that two spaces are better with monospaced fonts, like typewriters, but that one space is better with proportional fonts. If you want to test which spacing practice is superior, you must test using a proportional font and single line spacing, which is how the vast majority of real-world documents are published. Also, whether or not the paragraphs were left-aligned or justified might make a difference. Two spaces can result in very odd and disruptive spacing when the text is justified, but most professionally printed documents are indeed justified. The monitors used in the study were also of 2002 vintage, and hardly representative of the present-day, digital reading environment, much less that of printed works. In essence, the researchers were on a proverbial hunt for lost car keys under the street lamp because that’s where the light was best.
The researchers recognize this problem, but they hand wave it away with the unsubstantiated statement:
If the facilitation from two spaces is due in whole or in part to increasing the space relative to other spaces (e.g., to indicate not only the end of a word, but also the end of a sentence), then two spaces should facilitate reading even when text is presented in a proportional font where a single space is the same size regardless of whether it follows a punctuation mark or not.
But they present no evidence to support that belief, and in fact the evidence that the effect claimed by the study exists at all is equivocal, at best. A good study does not go places the data doesn’t.
What does the study actually show? First, there was no significant difference in reading comprehension for any of the four types of paragraph. In short, it appears that the number of spaces following a period or comma has no impact on reading comprehension. That’s a result that provides evidence that it may not really matter whether you use one or two spaces. It’s not definitive evidence, because the test is in no way similar to real-world reading conditions, but it is evidence nonetheless.
It’s with reading speed that the results get interesting. First, the study showed that for all readers, two spaces following commas significantly slowed reading speed. This isn’t terribly surprising or controversial—after all, no one advocates for two spaces following commas—but it does show that spacing can, at least in some circumstances, impact reading speed.
There was a small, that is around 3%, increase in reading speed for paragraphs that had two spaces after periods and one space after commas, but this effect only applied to the two-spacers. The one-spacers had no significant difference in reading speed between the four types of paragraphs. So it would seem, that if you are accustomed to typing two spaces after periods, you read paragraphs typed in this manner a bit faster. But if you are accustomed to typing only one space after a period, then there is no difference. This is an odd result and deserves further looking into, although the effect size is quite small, and it may be that when the experiment is replicated this effect disappears.
The eye-tracking also showed that readers tended to dwell longer on the spots where there was only one space following a period, which indicates that readers’ brains take longer to process those characters, but this effect did not impact overall reading speed. The conclusion that seems most likely is that two spaces following periods is easier to read, but this effect, while measurable, is insignificant for any practical purpose.
The really interesting result, though, is that the two-spacers had an overall faster reading speed than the one-spacers, regardless of the type of paragraph. Again, that’s a really surprising result and deserves further investigation. At the least, the experiment should be independently replicated to see if this result appears again.
So what do we have? The study, as written, is flawed. It’s primary conclusion is not supported by the results, and indeed the experimental design used cannot produce the data necessary to reach the conclusion that two spaces is better than one. Also, the description of the experimental design is missing critical parameters that are required for replication. But, nonetheless, it does produce some interesting results, even if those aren’t what it is touted to produce.
As far as I know, all the good evidence available, which includes some of the evidence in this study, indicates that there is no reason to believe that either one or two spaces after periods is superior in any measurable way. That is, it really makes no difference. Given that, the best practice is, therefore, to follow convention. And since, pretty much without exception, all professional typesetters and publishers use only one space after periods, that’s what you should do too, at least until someone comes up with a study that provides solid evidence to the contrary.
For further reading, I recommend Matthew Butterick’s article on this subject, which reaches pretty much the same conclusions that I did. His website is also a great resource for all thing typographical.
Sources:
Butterick, Matthew. “Are Two Spaces Better Than One? A Response To New Research.” Butterick’s Practical Typography, 30 April 2018, https://practicaltypography.com/are-two-spaces-better-than-one.html.
Johnson, Rebecca L., et al. “Are Two Spaces Better than One? The Effect of Spacing Following Periods and Commas During Reading.” Attention, Perception, and Psychophysics, published online 24 April 2018.
Selk, Avi. “One space between each sentence, they said. Science just proved them wrong.” Washington Post, 4 May 2018, https://www.washingtonpost.com/news/speaking-of-science/wp/2018/05/04/one-space-between-each-sentence-they-said-science-just-proved-them-wrong-2/.
