WebAIM - Web Accessibility In Mind

E-mail List Archives

Re: External Link Icons

for

From: Jukka K. Korpela
Date: Sep 29, 2007 5:30PM

Karl Groves wrote:

>> Karl Groves wrote:
>>
>>> I have witnessed, on numerous
>>> occasions, participants in the lab leave the site we were
>> testing and
>>> be completely unaware that they had done so.
>>
>> Such things do happen, though we cannot really know how often
>> - we can observe such situations but we don't know how
>> representative they are.
>
> We can assume with relative certainty that, provided a properly
> created set of user personas that, if participants fitting those
> personas have problems knowing they've left the site, others in that
> user population will have similar problems.

That assumption does not tell us how representative the situations are.

> This is entirely the
> point of qualitative usability studies!

If they are qualitative, then they don't (by definition) answer any "how
much" questions.

> So your point that "we don't
> know how representative they are" is wrong. We do know because the
> participants themselves are representative samples of that site's
> user population.

Representative samples? That would be rather extraordinary. Which _sampling_
method did you use? I'm afraid you are using just pseudoscientific jargon.

If you told _how many_ of the participants were unaware etc., then we might
try to deduce something quantitative on common sense basis, very roughly -
not statistically, since (to begin with) you have no representative sample
and even if it were, it would be far too small for any generalizability on
statistical grounds.

> But enough people are
> likely to have this problem that it is a very real concern.

You haven't given any quantitative information.

Moreover, you haven't proven that it is a problem at all.

> As I said in my last response, I agree with you on the premise that
> the site owner being concerned about "losing visitors" is a silly
> concern.

Yes, but that's the _real_ reason behind the desire to put External Link
Icons around.

> But I'm coming from the view that the user themselves may
> not want to leave and may inadvertently do so.

"May" is the operative word. That's just speculation for now. Besides, if
that _is_ a problem, then is there any reason to think that External Link
Icons would address it? If the problem is in the way of using the Web, then
the solution is to teach those people use it better, e.g. to find out ways
to recognize _in general_ where they are going etc.

>> If it _really_ matters that some links are external, you
>> normally can and should _say_ that they are external and why
>> that matters.
>
> Placing this notice within the text is not reliable for links which
> appear inline with other text.

Why wouldn't it be? If people miss a normal textual explanation, will they
even understand link texts? If I say
"There are more details on this in the Wikipedia article <a ...>...</a>,
which looked relatively accurate when checked on ... (but may have changed
of course)."
then I've expressed what needs to be said.

> I have seen users clicking links indiscriminately when lost
> on a site.

Would they be helped if some links had cryptic icons associated with them,
different in different sites?

Jukka K. Korpela ("Yucca")
http://www.cs.tut.fi/~jkorpela/