Embedded Links and Online Reading Accessibility: Whitney Quesenbery and Caroline Jarrett, #stc10
May 7th, 2010 | Posted in blog 16 Comments »
In this video, I talk with Whitney Quesenbery and Caroline Jarrett about the feasibility of removing links embedded directly within paragraphs — which Kathryn Summers and Ginny Redish describe as “exit points” that confuse and disorient low-literacy readers.
Sponsors
Tags: accessibility, caroline jarrett, Ginny Redish, kathryn summers, links, low-level literacy, reading, Screencasts, Web Design, 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.
3834 Subscribers
Greg DeVore (4)
Mark Baker (4)
corporate gifts (2)
Great Clips Coupo... (2)
mma (2)
Tammy P (2)
Childrens Charities (1)
Christine Astle (1)
The site Plainlanguage.gov is good, but when people ask me for writing tips, I always send them to the Center for Plain Language instead. I have added Plain Language Expert to my résumé since learning about the term a few months ago. I’ve been saying these things for years, but now I have a name for it.
On links within paragraphs, I am afraid that it’s too late to do it any other way. Whitney Quesenbery and Caroline Jarrett both make good points, but they are talking about people who are not very familiar with the medium of hypertext. I am afraid that hypertext demands short links embedded in paragraphs. I am going to have to think long and hard about this idea, but I don’t see it catching on.
Shawn, thanks for starting off the conversation thread on this post. I also like your tip about adding plain language expertise to your resume. Re the discussion about links within paragraphs, I agree with you there.
By the way, I subscribed to your blog just now, so I’ll be following your posts.
Shawn – welcome to the world of Plain Language.
I’ve been brooding on your point, which uses embedded links within the paragraph, and on Tom’s post, which also uses embedded links within the paragraph, and on the discussion.
You said “hypertext demands short links embedded in paragraphs”. I can’t agree with that. For example, the page I’m typing into right has lots and lots of links – very few of them are embedded within paragraphs. Most of them are split out: Tom’s list of recent posts, the names of people introducing the latest comments, navigation links and so on.
Tom’s post says this:
“In this video, I talk with Whitney Quesenbery and Caroline Jarrett about the feasibility of removing links embedded directly within paragraphs — which Kathryn Summers and Ginny Redish describe as “exit points” that confuse and disorient low-literacy readers”.
He’s linked three of the names (Whitney, myself, and Ginny) as embedded links. (Hi Tom. You can find Kathryn at http://iat.ubalt.edu/summers/)
He could equally well have put the links as a bulletted list at the end, like this:
- Whitney Quesenbery, WQusability
- Caroline Jarrett, Forms That Work
- Kathryn Summers, University of Baltimore
- Ginny Redish, Redish and Associates
This gives an uninterrupted reading experience and adds more information to the links.
For me (and I’m guessing, for you) reading this post on Tom’s site: It’s not a big deal. I’m a high-literacy reader with a great deal of familiarity with the topic I’m reading about, and I can easily make choices about whether or not to follow the links. And I think that most of Tom’s audience for this video cast are people just like me.
Kathryn was talking about writing for a much wider range of people, including the very large numbers of readers who do not read easily. They are less able to make judgements about following the links, and are much more likely to click on the first link that they come to.
So: if you’re writing for the general public, avoid embedded links.
And I have found that now I’m getting used to avoiding them, I use them less and less even in material for expert readers. Experts are likely to be busy and impatient readers – and therefore it’s best to avoid distractions for them, too.
Best
Caroline Jarrett
Caroline, thanks for the link to Kathryn Summer’s page. I updated the post. I admit that embedded links within paragraphs do serve as exit points and will probably confuse lower-level literacy readers. But I honestly don’t think it’s feasible to add a list of links at the bottom of the article. In the example you list, the list of links is a simple list. But what if I were referring to a variety of different posts, studies, ideas, and comments people made? If someone wanted to locate references to them, he or she would need to guess how the reference links related to the content in the article.
Alternatively, I would need to implement a reference system inline with the text, such as footnotes or parenthetical references with last names, e.g., (Johnson). The parenthetical references, similar to MLA style, would seem so unlike the rest of the web that it would draw attention to itself.
I was surprised during this presentation that more people didn’t object to this blanket statement about removing inline links. It may improve one’s ability to get through a post, and surely adding too many inline links is confusing, but in general the advice didn’t consider all the implications and consequences of adopting a link-less writing style.
First, Google will probably not change their search engine algorithm for this. Google is smart because it looks at hyperlinks in context. So if I create a link that says Caroline Jarrett’s book on form usability is helpful to anyone creating forms, that registers with Google differently than saying Caroline Jarrett’s Forms That Work is helpful to anyone creating forms. The first is optimized for the term “form usability,” while the second is optimized for “Forms That Work.” It makes a difference with SEO. I don’t think companies and individuals will wholesale sacrifice SEO in order to improve readability a little bit. It’s even more bold to think that Google itself will completely change the way it ranks and sorts information.
Second, if a user is trying to actually find more information about a topic you mention, the embedded link provides better context and placement for the reference. It makes sense to refer to the user to the content in the moment you’re addressing it rather than in a list of links at the end.
Anyway, I guess I have strong feelings about this topic. Hyperlinks in context are simply how the web seems to work. Thanks for letting me interview you at the conference about your reactions. Also, thanks for your followup comment.
Tom, why don’t you pick a few of your posts, and send them to Whitney to remake for you? I bet that she can make them more readable without giving up the SEO.
[I don't see why you'd need to change much in the wording of the link, anyway. I've been looking at some of your more complex posts, and I think I can see how they'd work just fine.]
But also, I think that part of the difference here is whether you’re talking about help or a blog. I mean, for your blog, you may care a lot more about SEO than I do in my documentation, where I have more of a captive audience.
Yes, I’d rather that my readers found me first when they Google for help. But on the other hand, if they go to my help, I’d rather that they understand it, read it easily, and don’t call us.
I think the discussion about embedded links is something that will live and die only within tech comm and usability circles. 99 percent of the web won’t care, and really I’m not persuaded to the model of extracting embedded links from content in the first place. Posts on this blog are intended for more literate users anyway, so it would mostly be help documentation that I would ever implement the model. Even then, I still routinely insert embedded links. Wherever a phrase in a help topic links to another topic, I try to add it, because these little cross referencing hyperlinks make information more findable within help.
By the way, Penelope Trunk’s blog would be a good case study for inline links. She includes more links within the sentences of her posts than almost any other blog I read.
This was a very interesting interview. Thanks for making it, Tom. I think it is important to remember the “it depends” remark. I, too, was a bit “oh, no” because I have taken to hyperlinking text like a fish takes to water.
However, I recalled a link procedure we had at one workplace. All links went at the end of the paragraph. Our manager was a real stickler for wording. Sometimes too much of a stickler we thought. You’d write an explanatory paragraph. Then, you’d have “For more information about (insert the what), see (insert the where of the cross-reference).” That was our text template, and it was always at the end of the paragraph. (Note: this was for datasheets that were delivered as PDFs. We did not produce HTML material, so all cross-references were somewhere else in the PDF, with an occasional URL reference.)
Whitney’s and Caroline’s argument meshes with this. Our style of links in paragraphs meshes with their proposal.
When you consider the audience, it also makes sense. Federal or public service sites must serve the general public, which means they must consider their language. It is also getting back to some good principles – get to the point first and give details later. Whitney gives an example of three choices where you could say it is imperative that the reader does read all three choices before traipsing off to explore any of the choices. You may also be familiar with Microsoft’s expand and collapse trick where you are presented with a complete paragraph that can stand alone. Some words are, however, marked in some way so you know you can click them and they expand with additional, explanatory text.
In other words, get the main point across, then deliver the extras.
It suddenly dawns on me that this is a variant on progressive disclosure. When you might overwhelm someone with too much info at one go, you layer info so you can reveal a little bit at a time.
The bottom line (IMHO) is that federal/public service sites must consider these issues about links. Technically, you don’t have to do anything on your site. You could have perfectly intelligent people read your site who do have certain learning difficulties and who would appreciate a less-linked article. There is a solution. Let people subscribe to post by email. There are post email programs that send text-only mails out, which mark each hyperlink with a footnote number and list the link at the end of the article. It works quite nicely.
Here’s another real story to add some weight to this argument about not embedding links in paragraphs of text.
The story is from some usability testing I did for a university client. The users were international prospective students, and the testing took place in Australia. All participants had English as a second language. All were in Australia doing either an IELTS (an acronym they are VERY familiar with) language course or an academic bridging course to prepare them for entry to university-level study. None were taking these courses at the university (or its associated college) whose site was being tested.
One key task we asked participants to do was find out if the university offered a course they were interested in. Part of the task was to find the entry requirements, fees and associated information.
When looking for entry requirements, participants knew they had to find 2 pieces of information – academic and English language skills. None of the participants (if I recall correctly, I think there were
found the language standards. ALL – every single one of them – found the right page, the right part of the page, and then followed an embedded link that was a mere 3 WORDS AWAY from the information they were looking for. None found the answer on the linked site, and none came back to the page they followed the link from. This happened across a range of courses, because all the course information pages were done in the same way.
The offending text and link went something like this: ‘Applicants need an IELTS score of 7 blah, blah …’. ‘IELTS’ was a link that took them to the IELTS website. They all clicked on it.
For me, there were two morals to this story:
1. Don’t link just because you can. You must have a very good reason for adding a link to text. Linking to the IELTS site was ridiculous. All the participants (and the broader target audience) would know what IELTS was by the time they were able to read the text. It reminded me of web designers who write that they do ‘HTML, CSS and blah, blah..’ and link the words ‘HTML’ and ‘CSS’ to the W3C spec pages. As if their clients want to read the specs (and as if many of the designers had)!
2. Don’t link a tantalising trigger word anywhere near key information.
I have a lot of trouble convincing people of this until I show them examples of how to get around it. Most people think you’re trying to tell them they cannot link at all. Showing them how to move the link(s) out of the way usually works.
(I hope this made sense – I hate writing in such tiny boxes.)
Dey –
Thank you for that research evidence.
Caroline and I have seen similar problems (with a similar audience) testing the Open University web site. The problem is not just tantalizing links, but language that may sound elegant, but is far from plain. Participants (even native English speakers) would be tripped up by sentences that hid the critical details in subordinate clauses and other problems.
On sites where it’s important that people get the facts, Horton had it right: “Say what you mean, and mean what you say.”
Karen Mardahl sent me a link to an article in Wired with an excerpt from Nicholas Carr’s new book. Here’s a few clips on neuroscience research that seems to support the idea that (except in some specific situation), using too many embedded links does not work well.
“By the end of the decade, the enthusiasm 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.””
“There’s nothing wrong with absorbing information quickly and in bits and pieces. We’ve always skimmed newspapers more than we’ve read them, and we routinely run our eyes over books and magazines to get the gist of a piece of writing and decide whether it warrants more thorough reading. The ability to scan and browse is as important as the ability to read deeply and think attentively. The problem is that skimming is becoming our dominant mode of thought. Once a means to an end, a way to identify information for further study, it’s becoming an end in itself—our preferred method of both learning and analysis.”
“Even though the World Wide Web has made hypertext ubiquitous and presumably less startling and unfamiliar, the cognitive problems remain. Research continues to show that people who read linear text comprehend more, remember more, and learn more than those who read text peppered with links.”
Whitney, thanks for sharing that quote. Interesting. It does seem to provide a stronger argument for removing embedded links. I agree that comprehension declines and users skim and skip around more when there are embedded links in the content. But you could equally argue that users are skimming and skipping because they’re looking for content that isn’t on the page. We skim and skip and click because the information we need isn’t on the page we’re reading, and we’re quickly trying to find it. In a world of ever-increasing information, we don’t have time to read a full page slowly and carefully. That full page may not contain the answers we need. Through the system of embedded links, we can more easily find the content we’re looking for.
Wow. I wasn’t suggesting that pages shouldn’t have links. Of course, the whole value of hypertext is that links can take you to other information, deeper discussions, or definitions (to name a few).
This discussion is about how these links should be placed within the text. The style of embedded links that seems to not work well is when a word or phrase is linked without any contextual cues to indicate what type of information that link points to.
Does a link on “breast cancer” in an article about whether obesity increases your risk of getting certain cancers go to a definition of breast cancer? To general information about that cancer? To additional statistical evidence? Without context in BOTH the surrounding sentence AND the link text itself, the reader has no way of knowing.
This is as much a problem of writing style as anything else, so who better to debate how we provide these hints than a bunch of people interested in writing and usability.
Here’s 3 examples:
Wikipedia is very consistent: all links within the paragraphs all go to other Wikipedia topics; links to external information are footnotes.
The New York Times is confusing: how would you know that the words “has been criticized here” in a sentence “David Cameron, the new prime minister, been criticized here for not standing up more forcefully to the United States” goes to a single article in the UK Telegraph, when the words “David Cameron” go to a biographical summary provided by the Times.
On the other hand, a link like “some of the [link] more common HAT features [link], such as single sourcing and print” suggests that it points to something like a table of common features. Instead, it points to a summary of points made in a presentation about the benefits of an HAT.
There is a usability effect of this confusion: On cancer.gov, most embedded links go to definitions. In usability tests, we see people clicking on these words without any clear idea of where the link will lead. Worse, once they learn that most of the links are to definitions, they stop clicking on links because they assume that all are.
Anyone who has had to learn to write to a specific citation style will appreciate the subtle changes in how you frame a sentence depending on whether you are using Harvard or APA (STC’s Tech Comm) vs ACS Style (IEEE Transactions on Professional Communication).
The shifts are similar in considering how you construct hypertext links. I would argue that if you are going to link words embedded within a larger paragraph, you have a responsibility as a writer to provide cues as to where those links (really) lead.
We are in the process of teaching ourselves how to write in a richly linked world.
Okay okay, you finally convinced me of this. It was the reading of a blog post that appears as a pingback in these comments that allowed me to truly see the increase in readability. I will experiment with more referenced footnote links.
[...] – Author Nicholas Carr: The Web Shatters Focus, Rewires BrainsTom Johnson – Embedded Links and Online Reading Accessibility: Whitney Quesenbery and Caroline…Twitter page for Cliff Tyllick, a Web development coordinator specializing in usability and [...]
[...] – Author Nicholas Carr: The Web Shatters Focus, Rewires BrainsTom Johnson – Embedded Links and Online Reading Accessibility: Whitney Quesenbery and Caroline…Twitter page for Cliff Tyllick, a Web development coordinator specializing in usability and [...]