Page 172 -

7.4 Goals of HCI case studies 159

asking additional questions of existing sources or consulting new data sources in
order to develop explanations that resolve the inconsistencies.
Case studies often draw upon many data sources. Documents, data archives, direct
observation, and participant observation (similar to ethnography—see Chapter 9) are
just a few of the possibilities (Yin, 2014).

7.3.4 EMPHASIS ON QUALITATIVE DATA AND ANALYSIS
The researchers were not specifically interested in measuring how quickly Sara com-
pleted various tasks, how many errors she made, or how quickly she learned to use an
interface. Case studies always contain a substantial qualitative component, focusing
on questions that help describe or explain behavior (Yin, 2014). In Sara's case, ques-
tions might have included “How did she use technology to achieve various goals?”
or “which tools did she use in a given circumstance?” The data needed to answer
questions such as these tends to be more qualitative than quantitative.
Case studies can certainly include quantitative components measuring traditional
metrics, such as task completion time, but these measures are not usually the sole focus
of the investigation. In Sara's case, the investigators might have measured the time it
took her to complete certain tasks or how frequently she used the tactile watch. As in-
teresting as these measurements might have been, they would not have been sufficient
to meet the goal of the study: a deeper understanding of her use of assistive technology.
The qualitative interviews, which provided room for in-depth discussion, elaboration
of concerns, and discussion of contextual issues, were crucial for achieving this goal.
Quantitative data might be used as a triangulation tool for corroborating results.
In this case, Sara's frequency of use of the tactile watch might be used to provide
supporting evidence for interview comments describing her perceptions of how she
uses the watch. If you choose to use quantitative data in a case study, you should be
acutely aware of its limitations: as your case or cases are unlikely to be representa-
tive of a larger class, statistical comparisons are generally not appropriate. Having
collected data on the frequency of Sara's use of the tactile watch over a period of
weeks, investigators might have sufficient data to investigate how Sara's use of the
watch changed over time, but they would not have been able to make any comparison
between Sara and other tactile watch wearers, or blind people in general.

7.4 GOALS OF HCI CASE STUDIES
Goals of traditional qualitative case studies generally fall into one of three categories
(Yin, 2014):

• exploration: understanding novel problems or situations, often with the hopes of
informing new designs;
• explanation: developing models that can be used to understand a context of
technology use; and