I was reading an article, and the author underlined a single word. Automatically I though 'this is a link, let's click it', but it was only a underlined word:

enter image description here

Is there a study that shows that users always think about links when they see an underlined word?

  • 1
    I assume you mean when reading on-line only? Though I think that it gets really interesting when you put Kindle or tablet-type devices into play. – Michael Kohne Nov 21 '14 at 18:53
  • Yes, I mean reading online – IAmJulianAcosta Nov 21 '14 at 19:27
  • Well, hyperlinking is the only excuse there is for using underlined text at all (at least for European scripts). So what else would users associate with it? – Crissov Nov 23 '14 at 7:38
  • 2
    @Crissov: Underlining was the equivalent of italicising, before wysiwyg computing. – MMacD Jan 16 '17 at 17:41
  • @MMacD Yes, underlining is adequate in handwriting and on mechanical typewriters etc., but who puts these on the web? WYSIWYG text editors should not even have that nasty U button (nor I and B in the way they do, but that’s another matter). – Crissov Jan 16 '17 at 19:59

It is hard to say whether everyone does, but the majority who use the interest would. Nielson Norman group guidelines specify underlining links and the majority of websites tend to follow this.

Nielson Norman Group - Guidelines for visualising links.

There is not much research on this so the best way to find out? Conduct a study yourself, you could have a paragraph of text on the screen with a few pieces of underlined text and could use heat maps to determine if people hover over the underlined text

Google did some research on the colour of links, they found that the colour was seen to be important: Google study

Some other things which are interesting just to think about: If you look at this answer, the links I have provided are not underlined. Something even more interesting (feel free to correct me if I am just missing something here), no underline option.

enter image description here


I've been surfing the net since it's inception. I can say that I perceive underlined text of the fore color to be plain text, and underlined; or not underlined text of a different color as possible links, and I usually hover over them if not underlined and observe the status bar to see if this is so. I also perceive text in a box as a button to a link and hover it as well looking for a tool tip etc.

I think most people would do the same, although I have read articles where a link was provided in a different color, not underlined and it was missed by its intended audience. Like anything else, there is no perfect way, but there are good standards available as mentioned by AEJBUG.

Following good standards, like following good language conventions improves communications and isn't that what the W3 is all about?


No, there wouldn't be any sort of study like that since underlining already had a meaning before the web came into being. Existing meanings tend to interfere with attempts to impose new meanings.

The traffic signal at the intersection of Thompkins St. & Milton Av. in Syracuse NY is "upside down" (green on top) and has been that way for ~100 years. The Irish residents of the area like it fine, but it's a little disconcerting to out-of-towners who have to take a quick decision about how to interpret what they're looking at. Do they obey the position or the color? Hard to tell, if there's no other traffic around to give them a hint.

Note that SE doesn't underline links, yet nobody is bothered or confused. Which we would be if we'd successfully been conditioned by the ukaz that links be underlined.

  • "No, there wouldn't be any sort of study like that" - I do not understand the reasoning put forth in that paragraph. "Note that SE doesn't underline links, yet nobody is bothered or confused." - the claim/question is whether an underline implies a link. No-one has claimed a link implies an underline. – O. R. Mapper Jan 17 '17 at 12:13
  • @O.R.Mapper: I agree that as a syllogism it doesn't work, but underline-means-link is only sometimes true in the context of the web and its byproducts, and definitely false in all other contexts. So the odds of some random underline, even on the web, hiding a link isn't that large. – MMacD Jan 17 '17 at 18:44
  • (continued) I suspect that if a study were done in which people were shown various bits of underlined text in a web context, with a small sum of money being the reward for identifying a given underline correctly (link or non-link) but a larger sum taken from them for getting it wrong, people would scrutinise the surrounding text pretty carefully each time before making their choice. – MMacD Jan 17 '17 at 18:45

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.