Database Reference
In-Depth Information
ily quantitative results of a laboratory experiment and in this they play an important
but secondary role.
Think-Aloud Protocol:
This technique, which involves encouraging participants to
speak their thoughts as they progress through the experiment, was introduced to the
human-computer-interaction community by [43]. Discussions about this protocol in
psychology date back to 1980 [19, 20, 21]. Like most methodologies, this one also
involves tradeoffs. While it gives the experimenter/observer the possibility of being
aware of the participants' thoughts, it is not natural for most people and can make a
participant feel awkward; thus, think aloud provides additional insight while also
reducing the realism of the study. However, the advantage for hearing about a partici-
pant's thoughts, plans, and frustrations frequently out-weigh the disadvantages and
this is a commonly used technique. Several variations have been introduced such as
'talk aloud' which asks a participant to more simply announce their actions rather
than their thoughts [21].
Collecting Participant Opinions:
Most laboratory experiments include some method
by which participant opinions and preferences are collected. This may take the form
of a simple questionnaire or perhaps semi-structured interviews. Most largely quanti-
tative studies such as laboratory experiments do ask these types of questions, often
partially quantifying the participant's response by such methods as using a Likert
scale [44]. A Likert scale asks a participant to rate their attitude according to degree.
For instance, instead of simply asking a participant, 'did you like it?' A Likert scale
might ask the participant to choose one of a range of answers 'strongly disliked,'
'disliked,' 'neutral,' 'liked,' or 'strongly liked.'
Summary of Nested Qualitative Methods:
The nested qualitative methods men-
tioned in this section may be commonplace to many readers. The point to be made
here is that in the small, that is as part of a laboratory experiment, inclusion of some
qualitative methods is not only commonplace, its value is well recognized. This type
of inclusion of qualitative approaches adds insight, explanations and new questions. It
also can help confirm results. For instance, if participants' opinions are in line with
quantitative measures - such as the fastest techniques being the most liked - this
confirms the interpretation of the fastest technique being the right one to chose. How-
ever, if they contradict - such as the fastest techniques not being preferred - interest-
ing questions are raised including questioning the notion that fastest is always best.
5.2.2 Inspection Evaluation Methods
We include a discussion of inspection methods because, while they are not studies per
se, they are useful, readily available, and relatively inexpensive evaluation approaches.
The common approach is to use a set of heuristics as a method of focusing attention
on important aspects of the software - interface or visualization - which need to be
considered [54]. These heuristics or guidelines can be developed by experts or from
the writings of experts. Ideally, such an inspection would be conducted by individual
experts or even a group of experts. However, it has been shown that in practice, that a
good set of heuristics can still be effective in application if a few, such as three or
four, different people apply them [54]. For information visualization it is important to
consider exactly what visualization aspects a given set of heuristics will shed light on.
Search WWH ::
Custom Search