How do you know which UX research method to
do when? Let's look at a few criteria to help us decide. First, think about watching or asking. Research can be either behavioral and observed
by researchers or self-reported by the user, which is also known as attitudinal. When your research question has to do with
what people actually do, or whether something is discoverable, findable, understandable,
or usable, it's most informative to observe this. So use a method like a usability test. If you're in very early design phases, like
you have sketches or wireframes, you can still observe usage by giving people tasks to do,
with just those paper sketches. As your design progresses, create high-fidelity,
dynamic, clickable prototypes to test. When you have design concepts, it's tempting
to just show it to users and ask what they think. But instead, direct the user to try and actually
do something. If you're wondering what people have to say,
like whether they believe they'll like or dislike something, that can be researched
with self-reported methods like interviews, surveys, and focus groups. Another criteria to help us decide which research
method to use when relates to numbers or stories. Or quantitative or qualitative research. If your goal is to compare products or designs,
to get benchmarks to which the future design will be compared, or to compute expected cost
savings from design changes, you will use a quantitative research method. Some quantitative methods to use when you
have these goals are card sorting, which calls for about 20 users, tree testing, which calls
for between 50 and 100 people, eyetracking heatmaps, which need 40 users, quantitative
usability tests, during which we watch people do tasks and take measurements,
like time on tasks and success rate, we start with about 20 users for these and the number
increases depending on the desired confidence interval and statistical significance; web analytics,
like A/B testing or multivariate testing, which require many more users; surveys always
require many more respondents. On the other side is qualitative testing that
we use to quickly and cheaply learn so we can iterate the design and improve it as much
as possible before releasing it. Very early prototypes can be usability tested
with just 2 target users. Make a change based on what we observe, and
then test it again with two more people. The next criteria is tangible or not. We also study people when we have no design
at all. Especially in very early design phases, we
want to learn how people work and live. A field study is a great method to use for
these needs, when we go to users and observe their every day activities. Then if we care about user behavior, or ease
of use for that matter, over time, then we want to do a longer study, or a longitudinal
study. This might be a diary study in which people
keep a journal or answer a series of questions; another option is to do a series of surveys;
and yet another appropriate research method is a video journal or camera study, where
people record themselves at certain times. The last criteria I'll discuss for choosing
UX research methods is context. If you desire situational data that comes
at the user's location, there are several options. One is a field study where you go to the user's
office, home, car, or other site. Then you can still get some context information
with the following methods, like a remote usability test. Even a live, online intercept study gives
a small amount of context and that catches the user while he or she is actually attempting
a task on your website. And another type of study that goes by the
same name, researchers intercept a person live at a place like a coffee shop or mall. These studies are usually very short and are
okay to do in a chaotic setting. The flip side of that are lab studies, where
you opt for a controlled environment so you can make the absolute most of the precious
time that you have with your user.
