On Sep 2, 2009, at 10:45 AM, Audrey Crane wrote:
Of course people may have more of a vague sense than a clearly-formed
question, but I'm going to ponder how to test this. I'm really
curious to try to manifest a response to having the page title or not
in usability. (Any thoughts?)
Audrey,
If, by "in usability", you mean "in a usability study", I can help.
(If that's not what you meant, then I don't know what you're referring
to, since usability is an adjective.)
You can't use a traditional formative usability study, since you can't
control the need for breadcrumbs. You have to construct a study that's
more analytical, that compares treatment options in a controlled
fashion.
To study something like the effectiveness of a specific design
treatment of breadcrumbs, you first need to understand what your
behavioral objective for the breadcrumbs are. What behaviors are you
trying to elicit?
For example, If the behavior is mental model development (in other
words, the users have a better idea of where they are within the
structure of the site), then you can show people sample pages with the
breadcrumbs and ask them to draw a diagram of the structure of the
site. (If you want to do this as a controlled study, then you can
repeat the activity by showing the control group pages without the
breadcrumbs and see if their diagrams are different. And if you really
want to get all study-crazy, you can then followup with a task-driven
study and see if people who were exposed to the breadcrumbs perform
better on the site than people who weren't.)
Another example: If the behavior is error recovery (aka I-don't-know-
how-I-ended-up-here-and-need-to-get-out), then you can drop people
onto random pages (the way they might if they clicked on a link in a
search engine) and ask them how they'd navigate to a target page. You
could see if the breadcrumbs get them anywhere useful.
In either case, showing the participants breadcrumbs with or without
the titles in a controlled fashion (either a within- or between-
subjects study would probably work fine), would tell you which
treatment performed better.
If you have other behavioral objectives, then, depending on what you
want people to do with the information, it would be fairly easy to
design the study.
Of course, these would not be cheap studies to execute. You'll need a
lot of participants to control for the various interfering variables
involved (domain knowledge, tool knowledge, experience with
technology, education level, performance anxiety issue, and others).
It'll take a decent stats person to clean up the data and report any
conclusive results.
And here's the kicker: if your results are like our results, you'll
find that virtually nothing you do with breadcrumbs will make a
difference. So, I'm predicting that after all that effort, you'll find
that you've not discovered any benefit to either treatment.
Our studies show that people don't form any better mental models about
the site when they encounter breadcrumbs than when they don't. (Our
studies also show that users don't need a representative mental model
of the site structure to successfully complete tasks on a site, so
breadcrumbs aren't really solving a problem here.)
Our studies also show that breadcrumbs are not the best treatment for
error recovery. More explicit links (the best being 7 to 12 words in
length) work significantly better. Of course, you're still treating
the symptom. It's better to solve the problem and prevent the user
from needing to recover from an error.
Hope that helps,
________________________________________________________________
Welcome to the Interaction Design Association (IxDA)!
To post to this list ....... [email protected]
Unsubscribe ................ http://www.ixda.org/unsubscribe
List Guidelines ............ http://www.ixda.org/guidelines
List Help .................. http://www.ixda.org/help
