Personal Attribute Judgments - Modeling Users' Experiences with Interactive Systems - page 35

Fig. 2.8 Importance rankings for reliability attributes for designers and users. Only 2 design-
ers reported reliability concerns in contrast to 7 users. Five users ranked reliability attributes
within their three most important attributes.

7th most important attributes while five users ranked reliability within their three
most important concerns. Hence, although most users prefer TouchToPrint, they
have some concerns that can potentially turn into failures to motivate use.

2.7 Discussion

In this chapter a simple and effective technique was presented for acquiring diverse
views on Repertory Grid data. The analysis revealed systematic differences between
the product qualities that users and designers appreciated within this limited set of
products. The proposed analysis procedure, however, relies on a number of assump-
tions.
First, it assumes the individual as a homogeneous entity. Attribute judgments of
a single individual are assumed to be well predicted by a single two-dimensional
configuration, and the emphasis lies in identifying individuals in a sample that form
an approximately homogeneous user group. At certain occasions this might be a de-
sired property of the technique, e.g. when the researcher wants to characterize single
individuals such as identifying differences in the perceptions of different stakehold-
ers in a design team. Diversity could however exist not only across different in-
dividuals, but also within a single individual, in the sense that different attribute
judgments of a participant may reveal different, complementary, views.
Second, diversity across individuals was explored using a single measure, that
of overall dissimilarity between the products. This approach assumes that individ-
uals are able to consistently summarize all attribute judgments in a single rating of
dissimilarity, an assumption that is ubiquitous in the field of psychometrics were
dissimilarity serves to summarize various perceptual judgments such as the per-
ceived blur and noise in an image (Martens, 2003). One might, however, question
the degree to which this holds in more complex settings where various cognitive
judgments are made for the quality of interactive products.