Removing Inline Links to Increase Readability
June 19th, 2010 | Posted in blog 28 Comments »
In the unfolding saga of inline links within posts and the decline in readability that these links bring about, Adriel Hampton’s post helped me persuade me more to this idea. Hamptom quotes from Nicholas Carr’s book, The Shallows: What the Internet Is Doing To Our Brains. Carr writes,
(In a 2001 study) one group read (a short story) in a traditional linear-text format; they’d read a passage and click the word next to move ahead. A second group read a version in which they had to click on highlighted words in the text to move ahead. It took the hypertext readers longer to read the document, and they were seven times more likely to say they found it confusing. Another researcher, Erping Zhu, had people read a passage of digital prose but varied the number of links appearing in it. She then gave the readers a multiple-choice quiz and had them write a summary of what they had read. She found that comprehension declined as the number of links increased—whether or not people clicked on them. After all, whenever a link appears, your brain has to at least make the choice not to click, which is itself distracting. … A 2007 scholarly review of hypertext experiments concluded that jumping between digital documents impedes understanding.
In other words, the more hyperlinks that you embed within your sentences, the less readable your posts become because the brain must make a decision with each link whether to click it for more information or keep reading. After several of these links, your brain starts to take on more cognitive load. As a result, it’s easier to get sidetracked with tangents or to lose retention of the content.
Every time your readers see a hyperlink in your text, they have to pause and ask themselves whether they should click that link and follow that path, or just stay the course ahead.
For a more in-depth reading of Carr’s argument, see Carr’s article in Wired, “The Web Shatters Focus, Rewires Brains.” In that Wired article, Carr explains,
By the end of the decade, the enthusiasm [for hyperlinked text on the web] was turning to skepticism. Research was painting a fuller, very different picture of the cognitive effects of hypertext. Navigating linked documents, it turned out, entails a lot of mental calisthenics—evaluating hyperlinks, deciding whether to click, adjusting to different formats—that are extraneous to the process of reading. Because it disrupts concentration, such activity weakens comprehension. A 1989 study showed that readers tended just to click around aimlessly when reading something that included hypertext links to other selected pieces of information. A 1990 experiment revealed that some “could not remember what they had and had not read.”
To paraphrase, because of over-linked text on the web, our reading habits have become more shallow. When reading online, we skip and skim. We read a bit and click a link, and read some more and click a link. This surfing and browsing results in a shallow reading experience.
The decline in comprehension presents one of the paradoxes of the Internet: Although the Internet presents us with vast amounts of useful, enriching information, at the same time it also shortens our attention span, reduces our comprehension abilities, converts us into shallow readers, and weakens the intelligence we have cultivated.
Since the STC Summit in Dallas, I’ve had heated discussions on this site about inline links, and Whitney Quesenbery and Caroline Jarrett have tried to help me see things another way. But it wasn’t until reading Hampton’s post, which omits inline links, that I started to see the improvement in readability that results when you remove the inline links. Stripping away all those inline links really did help me focus on the content.
Though I was adamantly opposed to the denunciation of hyperlinking on the web, and I fought against the abrupt and unargued dismissal of inline links during Kathyryn Summers’ and Ginny Redish’s Summit presentations, I am now coming around to a new point of view.
My only issue is in finding the best way to correlate the endnote links with the reference points the post. It’s confusing to guess how the references match up with the sentences in the post. For example, in Hampton’s post, my video interview of Quesenbery and Jarrett is listed at the end, but I’m not mentioned by name in the post. So is the list of links at the end a bibliography of suggested reading? Am I one of the “some bloggers” reference in the penultimate paragraph? Or are the endnote links more like related posts?
It’s not that difficult to correlate the endnote links with the post content. Having taught composition in college for four years, I know how to make a general References list — you just match up the author’s last names in the References list with last names in your sentences. But I’m guessing that doing it in a consistent and detailed way will be more tedious and require more effort than most people are willing to exert. Still, I’m going to give it a try in my posts for a while.
References Cited
Carr, Nicholas. “Author Nicholas Carr: The Web Shatters Focus, Rewires Brains.” Wired. May 24, 2010.
Carr, Nicholas. The Shallows: What the Internet Is Doing to Our Brains. W.W. Norton & Co., 2010.
Hampton, Adriel. “Usability: Are Your Hyperlinks Destroying Your Readers’ Brains?“ Gov 2.0 Radio. June 16, 2010.
Quesenbery, Whitney, and Jarrett, Caroline. “Embedded Links and Online Reading Accessibility: Whitney Quesenbery and Caroline Jarrett, #stc10.” Idratherbewriting.com. May 7, 2010.
Photo by nozoomii on Flickr.
Author Home Pages
Not cited, but useful for reading more about the authors cited in this post:
Sponsors
Tags: caroline jarrett, distractions, featured, Ginny Redish, inline links, internet, nicholas car, readability, Web 2.0, whitney quesenbery
If you liked this post, keep updated with new content: Subscribe to I'd Rather Be Writing.
Both comments and pings are currently closed.
3967 Subscribers
Jen (3)
DiSc (2)
Jenny (2)
Aneesha (1)
Barrie Byron (1)
Catherine Crane (1)
David Farbey (1)
Dawn (1)
Diane (1)
Fer O'Neil (1)
I’ve seen some blogs use superscript numbers at the end of a sentence, and then use a numbered list at the bottom of the post. I agree that using links in a sentence does cut down on the readability, but I don’t think a consistent method has emerged.
What’s breaks down readability even more is the use of ad links in sentences. Quite often I’m browsing a site, and the cursor moves over one of these links and then an ad pops up.
Tom, this explains something for me. For the last couple of years, I have developed a way of surfing the web that works for me. When my browser opens, with a couple of clicks I open about 35 web sites. Most of them are discussion groups or news pages. I right-click on the messages and headlines that look interesting and let them open in new tabs (but don’t read them yet).
Without being fully aware of it, this practices and others separates the experience of (a) skimming and selecting what to read from (b) actually reading for understanding. Your article has confirmed what I suspected; that it’s difficult to do both at the same time time.
Tom, wouldn’t it depend on what kind of writing we were talking about? In a blog post, which is basically an essay intended to persuade or inform, the links are a distraction. But in a piece of technical documentation — say, a set of steps — links can be very helpful. They can point a novice user to more detailed information, without distracting an experienced user.
Another thing: not everyone accepts the assertions that Carr makes in “Shallows.” Here’s another viewpoint (NY Times book review).
You’re right about the danger of applying this same model to instructional material. Inline links in a html-based help can provide a strong way of navigating among various help topics when users are looking for information.
Tom,
Like you, I’ve reluctantly come around to the view that “blue measles” — those spots that inline links create in a chunk of text — are usually best avoided. I agree with Larry Kunz, though, that it depends. In some contexts, it’s a service to put the link right where the info is needed.
As to your experiment, it would be helpful to add one little word before your “References” heading: “Cited.”
FWIW, way back when cross-references were the only links available, I was disappointed to read studies showing that people rarely follow those pointers even when the related info is critical to the task. Over the years I’ve greatly reduced my use of cross-refs and links.
Marcia, thanks for the tip on changing References to References Cited. I updated the post.
Hi Tom!
Over time, I have noticed a decline in my own reading retention. I read frequently, but most of that reading is online where there are tons of inline links either offering more information, or trying to sell something through an ad popup.
I think superscripts might be helpful and less distracting than inline links; however, your clear writing in this blog post made it easy to figure out which references go with which quotes. I too think it depends.
D
Many of the assertions in Carr’s essay and book about the decline in our reading patience and retention are hard to argue with. Re the superscript idea, I might try that. There are WordPress plugins that facilitate superscript footnotes.
I could agree, to a limited extent. One thing that you loose with pushing all your links to the bottom (this post is a perfect example) is context. After finishing the “content” of the post, I have to scroll back and forth to mentally associate the links with the content, to figure it if it was really a link I wanted to follow.
I see this as an area where sites could do better, and people should probably experiment with. I don’t think that rushing to the conclusion that we should treat the web just like print is a guaranteed “best way” to make things better.
After going back and re-reading your post I found something else that bothered me. Reading the parts where you cite a studies, I was fulling expecting a link to those studies. The post is long enough that I’d have to scroll down quite a bit to get to the footer notes. This left me with the impression while reading the post that you simply didn’t link to those studies (which would be very bad form for information on the web). One experiment that might be helpful here is a simple toggle option for the inline links. It could default to off, then I can turn it on and skim through the content again to see what items I want to follow up on (in their proper context, not as footer items).
Thanks for your comment. The link toggle is an interesting idea. I’ve never seen it implemented, though, so it might throw readers off by the mere fact that it’s unconventional. The superscript numbers referring to footnotes, like Wikipedia does it, might be a better solution. It still presents inline links, but the links wouldn’t jump directly to another site. I could also style the superscript references so they didn’t jump out as links.
Superscript numbers are more like “after line” links rather than inline. At least it’s my understanding of old-school (print) footnotes that they’re best put at the ends of sentences rather that directly following mid-sentence words. This approach makes for less disruptive reading.
I like the idea of using the superscript numbers as text, not links. Avoids the distraction of the “blue measles” effect as well as avoiding the uncertainty (will this take me to another site?). If the numbers are unlinked, people can more easily ignore them, while the few readers who care have what they need to dig into the references.
At some extent I agree with you. Hyperlinks (if too many) can distract me in reading one article. But, I don’t like the idea of using superscript number too. Oldfashioned and doesn’t seem practical. Anyway, hyperlink once was the answer to cites or footer note (I want to know more info concerning the post RIGHT AWAY. Not have to scroll down).
So, hyperlink then. Just don’t use it too much.
Cheers,
Pam
I think the real issue is that consuming content from websites rarely resembles traditional print-based reading. One of the more interesting usability metrics is eye tracking, which shows users generally scan very quickly for navigational markers based and then “read” the web content in non-linear ways—assuming it’s text and not visual media. One of the more interesting recent findings is that many webpage readers are actually adapting their reading strategies to accommodate the constant insertion of advertisement—so, many users now ignore “traditional” ad areas…and so the advertisers adapt as well.
Hyperlinks add another qualitative difference to the overall user experience and medium and make it even more difficult to really compare web-based reading to print-based reading. Factor in new online publication “experiences” like Wired and their tight integration of video, 3D, and augmented reality into the mix and you’re forced to step back and ponder on more creative integration of visual and textual media into a single publication.
Because of the different ergonomic factors (these still do play into the process when comparing different reading media), user motives, demographic/age preferences and so many other variables, I find it difficult to really look for any substantial equivalency between print-based reading and web-based “scanning” and come up with a strong evaluative assertion that ranks one over the other.
I think we often underestimate the cognitive differences that underlie web-based as opposed to print-based “reading.” We often associate (for valid reasons) print-based reading with deeper, more focused content acquisition and web-based reading with rapid content uptake. The underlying cognitive activity driving each activity is different. Where being more deeply focused on concepts and the sequence of a narrative might be helpful for a print-based article/story, this attention to detail would be counterproductive for the more fluid environment that is the web, which is also quickly evolving and therefore makes it hard to evaluate as a physiological/cognitive process—let alone compare to another type of reading. I say all of this simply to make the point that I found his thesis somewhat problematic, “Although the Internet presents us with vast amounts of useful, enriching information, at the same time it also shortens our attention span, reduces our comprehension abilities, converts us into shallow readers, and weakens the intelligence we have cultivated.”
-Chris
Definitely Carr’s prose is a bit overboard. But the basic argument IS very convincing if you’re a writer wanting the reader to actually focus on the information you’re trying to convey. Happy middle somewhere?
My basic web experience it based on “right click, open in background tab”.
I like to open inline links that I might be interested in, in the background, but continue on reading the article.
Perhaps another method would be to add the links to a “wish list” or read later list with javascript….
Tom – the reason your video is mentioned in Hampton’s post is explained in the link that follows it – someone suggested it to him, so he included it as a reference.
I think a lot of this discussion is about extremes and not about moderation. Who is your audience and where are you writing. Don’t forget the old chestnut “it depends”!
For technical documents, I’ve learned to include links at the end of a paragraph. The text for the links there is: “For more information (about X), see (cross-reference in same body of info or external resource). I’ve seen Danish newspapers do something similar lately. An article can have 20 short paragraphs, for example. Now and then, a paragraph is following by a URL to some related article the journalist thought might grab your attention. At one level, those links annoy me. However, they are very easy to skip and I do not feel my reading is disrupted at all. (Here’s an example with just one link in the middle of the article: http://politiken.dk/udland/article1005401.ece)
One thing that must be remember throughout this discussion is meaningful link info for screen readers. Screen reader users have always been able to read a list of all the links in one article. When the description of it (the part between the a tag) is only “click here” or “read more”, the links are useless. Meaningful info will also help the sighted reader with this method of putting links in one tidy place:
1. article about this bit
2. website for so-and-so
3. etc.
I tend to do what Bill Creswell mentions – open a tempting link in a background tab and continue my reading. Also, I might open and read something that provides background info (e.g. “as mentioned in my previous article”) to see whether it is important to read first.
Thanks for not using embedded links. I find that I am also side-tracked by the advertisements in the top-right corner of the page, especially when they are animated and keep changing.
Tom, I don’t find footnotes/endnotes distracting in online articles if they’re implemented the way programs like Microsoft Word implement them – a superscript number.
Here’s a WordPress plugin that allows you to do that easily: http://www.elvery.net/drzax/wordpress-footnotes-plugin.
I think this discussion is interesting, but it seems to be author-centric to a fault. In a medium where the cost of switching is so low and the potential value of switching is so high, an author-centric approach to publishing is only going to work for the very, very best authors — we’re talking 1 in 1,000 or less.
For most of us, if we’re trying to gain readers, we need to take a user-centric approach. Do my users want inline links? 99% of the time, the answer is “Yes, my users are arriving via deeplinks from a search engine, and they want a page built for skimming with inline links.” Many of us think this isn’t true, or don’t want it to be true, but if we’re trying to gain readers, it is.
Larry Kunz’s point of the type of writing is key. Technical documentation requires many more links than essays or opinions.
But regarding links in any type of writing, here are two more points that have not been mentioned:
1)
Links obviously add value by providing context without clutter, by avoiding repetition, and by reducing the trip length to related info.
2)
As readers, we are still learning to deal with the added dimension of links. As we get used to reading documents with them, we will learn to ignore them (or not) automatically, as we do with noise around us when reading. Of course, links aren’t noise — they are shortcuts to relevant information.
There is no great concurrence between learning and wisdom.
Francis Bacon from:hypequote
I’m really surprised at the controversy behind this article. In my observations of WWW content, what works best for me when reading is for the author to embed links inside a sentence primarily to refer the reader to *parenthetical* information, like using literal parentheses. (This article could also be directed at the use of actual parenthetical phrases in prose.) If the author wants to direct me to click on a link to continue the content’s main train of exposition, then that link should be explicitly called out as such in the text.
Tom, did you know you’re probably the blog I link to most from mine?
I’m going to be posting a webinar I just ran with DCL and I’m going to be quoting you again linking to this post. I am delighted to see you’ve come over from the dark side! Inline links are a crutch and side-step the need for good information design or content strategy (not sure you’re a fan of that term?). Bah, I say. Bah and humbug!
PS – something that I didn’t see mentioned was the possibiity of simple side-notes. These make the context and scope of the link far more obvoius yet don’t break the flow as you’re reading Your eye need only jump across if your natural flow is already broken by expecting more information or clarification. This is more usual in print where ‘box-outs’ and such add content relevant to the section or para, but this is one time when coding it is harder for web, so the practice isn’t really implemented.
Noz, thanks for the note you sent about the comment marketing. I forgot to respond to your email about that.
Re inline links, I have mixed feelings them. Initially I was opposed, and then I saw the point of removing them, but when I tried to put the no-inline-links model into practice, it became clear that removing inline links was impractical. So I returned to my normal ways.
I think there’s a possibility of going overboard with inline links, but really, anyone who asserts that web content should be de-linked doesn’t publish much web content, since there’s no practical workaround for inline links. I tried endnotes, footnotes — they don’t work, and my readers hated them. The best practice is probably just to not overdo links.
Before you make a case for removing inline links, try it on a few posts to see if you can find an alternative method that truly works. I couldn’t. Any alternative method defies web conventions, so the alternative method becomes more distracting than inline links themselves.
The only option I seem to have is to reply to my reply?
Anyway – this is interesting. We’re discussing inline links out of context I guess. I would agree it’s probably ambitious to tar all in-line links the same. I’m speaking from my own point of view (as I’m wont) which is one who is helping people post technical and reference content online.
Consuming content for pleasure or casual learning online is a -different use-case. I would say if I was reading your blog, then I might want some inline links (but I’m still going to take up your challenge of working without them), BUT if I was accessing your blog to look up a quick fact or follow a procedure, then they would take away more than they gave.
So I was over-generalising without qualificaiton, which is wrong wrong wrong, but I will stand by the sentiment as it applies to technical communication.
I set the comment nesting only to 2 levels. I just increased it to 4, but I’m not sure I’ve styled it in my theme. You’re right in that I’m mostly referring to web articles, not help. Maybe it makes more sense in help content.