Have you ever had to decide how wide a column of text you should use on a screen? Should you use a wide column with 100 characters per line? or a short column with 50 characters per line?
It turns out that the answer depends on whether you want people to read faster or whether you want them to like the page!
Research (see reference below) demonstrates that 100 characters per line is the optimal length for on-screen reading speed; but it’s not what people prefer. People read faster with longer line lengths (100 characters per line), but they prefer a short or medium line length (45 to 72 characters per line). In the example above from the New York Times Reader, the line length averages 39 characters per line.
The research also shows that people can read one single wide column faster than multiple columns, but they prefer multiple columns (like the New York Times Reader above).
So if you ask people which they prefer they will say multiple columns with short line lengths. Interestingly, if you ask them which they read faster, they will insist it is also the multiple columns with short line lengths, even though the data shows otherwise.
It’s a quandary: Do you give people what they prefer or go against their own preference and intuition, knowing that they will read faster if you use a longer line length and one column?
What would you do?
Dyson, M.C. (2004). “How Physical Text Layout Affects Reading from Screen.” Behavior & Information Technology, 23(6), pp. 377-393.
——————————————–
Did you find this post interesting? If so, please consider adding your comment, subscribing to the blog via RSS, signing up for our email list, and/or sharing the post.
Hello, Susan! First of all let me say that I'm an assiduous portuguese reader of your blog and that I translate, adapt and comment some of your posts to portuguese in my Blog!
My answer to this question reminds me the teachings of another guru of accessibility/usability (Jakob Nielsen) that tells us that we should give more importance to what user says rather than what the users do.
However, the answer is not so simple. According to my superficial reading of the paper, the study does not quantify the gains in speed against the number of characters per line and the understanding of the text. I mean that the gains may not justify the final appearance of the page (less appealing to users). More important than the number of characters per line is the text being adapted for the Web. Nobody likes to read large portions of text on the screen. If your goal is to get users to read a novel or other large literature on screen, then you should use the proven techniques that make easier the reading.
Yes, thanks for your comments. I agree that people don't like to read on a screen and there are things you need to do to make text more readable (another one of the 100 things that I'll get to in one of these blogs). We do, however, have to deal with this issue of preference vs. performance in many areas, line length being one of them. It's a philosophical decision we have to continualy make I believe.
Just to say that my previous message had an error! When I say "we should give more importance to what user says rather than what the users do" I meant "we should give more importance to what users DO rather than what the users SAY".
Hi Susan,
the columns form the Times in the picture you use are set as justify. I the paper you refer to, they used a regular left alignment. This leads to differences in line length throughout the text and a constant need to readjust. This could very possibly be slowing the reading and increase the memory load (as I would argue). This in turn might also lead to a worse comprehension performance (see Kintsch and Van Dijks model of text comprehension: Kintsch, Walter; van Dijk, Teun A. (1978) “Toward a model of text comprehension and production.” Psychological Review. Vol 85(5), 363-394.).
Since there are JavaScripts that enable hyphenation for websites (http://code.google.com/p/hyphenator/), this is definitely something that should be tested before inferring, that columns in general are less efficient to display text on screen.
Best regards,
Frederik
Frederik,
Yes you are right about the justification, however, that is an artifact here… The NYTimes Reader was not used in the study. I just used that as a picture in the blog (I guess I should have looked for an example where the text was not justified!).
Hi Susan,
This is so true! I find myself reading words by words literally when the column width is narrow while I could skim the line and just read the main words when the line is long such as arcoss the whole page.
I wish newspaper publishers know about this research :-)
Granted we are almost 12 months down the track since you wrote this, and a lot has changed in online media in this time, but, to the best of my knowledge, I am yet to come across an online site that actually uses columns. If the NYT used to, they don’t appear to anymore.
All blogging platforms tend to use ~100 characters per line as the default, including this blog :)
Is this a sign of where the online world has settled on this debate?
In terms of hardcopy newspapers, I think design / layout / spacing issues would mandate the use of columns.
Between 2 or 3 alphabet length is rhe rule for screen typography. I like to use 50 and 80 characters per line with a bigger fonts.
It’s indeed an interesting fact.
Maybe the best what you can do is to start with shorter lines to feed the satisfaction of the user.
After a while you jump to longer lines to serve better the reading task.
Quite useful article More important than the number of characters per line is the text being adapted for the Web.
I can’t help but wonder if we should really be concerned about how fast a reader is reading.
I’m not sure why the importance of a reader’s speed would matter if the reader was happy with the way the page was laid out and, even more importantly, if the reader was happy with the actual content itself.
If a reader likes the layout… and he likes the content… and he finds that content relevant and interesting, wouldn’t it be likely that he’ll continue to read?
I feel like I may be missing a major point here, so I’m hoping someone can shed light on why we should be so concerned about site visitors’ reading speeds.
Erin, I agree with you. Reading speed is only one measure and i may not be that important a measure. But people do talk about it sometimes, so I think it’s a good idea to make sure that people are clear about what the research actually says.
Susan, the 100 things articles are a great read. The clear language helps me better explain some of the design decisions I make, so thank you!
Reader preferences, reading speed, and reader satisfaction are all mentioned here, but I’m curious about the relationship of line length to reading comprehension. I’ll look around for info myself, but any comments or pointers to articles in that vein are most welcome.
Sarahjane,
All the research I’ve seen on this topic always refers to reading speed. The only research in regards to comprehension I’ve seen found that line length didn’t affect comprehension.
I think that the key point is “comprehension” because we read not to read, but to understand.
When I design a page I don’t care about the line length in itself, but about this in the relationship with the possible quickly comprehension.
Ex.
Can I eat? Can I have sex with it? Will it kill me?
have a slower understanding than:
Can I eat?
Can I have sex with it?
Will it kill me?
Regards
Nelson
I read your article about line lengths, but long lines like your recommendation (100 letters) are hard to follow going line to line vertically. Also, if you have to move your eyes too much laterally, it gets tiring very quickly. The 66 character line is closer to a paperback that has evolved from a long time of reader’s preferences.
I love your books! Keep ’em coming!
I wonder if when the tested the difference in reading speeds, if or how they tested reading comprehension.
******yes I know my grammar kinda sucked on that but you get it.