Information Technology Reference
In-Depth Information
analysis procedure: a) that all attributes of an individual may be analyzed by a sin-
gle two-dimensional configuration, and b) that dissimilarity ratings may be used to
model all attributes of an individual.
The procedure proposed in the next chapter explicitly aims at identifying whether
an individual's attributes can be modeled in one or two diverse views. Table 2.5
depicts the number of attributes being adequately modeled by the first and (in some
cases) a second view. Note that only two individuals required a second view while
for the majority of the individuals a single two dimensional view was sufficient
for modeling the majority of his/her attributes. The attributes of four individuals
resulted in no satisfactory MDS solution suggesting that their data are too noisy to
be analyzed in a meaningful way.
Figure 2.9 depicts the similarity of the resulting individual views. Note that the
differences between designers' and users' perceptions emerge also in the analysis
of individual attributes. Two views, one representing the designers group and one
representing the users group were derived from the configurations (see chapter 3 for
the exact procedure). The resulting two views were able to adequately model a total
85 attributes (56%).
2.8
Conclusion
This chapter highlighted two issues in the use of standardized psychometric scales in
measuring users' experiences with products. Firstly, a-priori defined measurement
scales are inherently limited in missing potentially relevant measures that the re-
searchers did not consider as relevant to a given context. Secondly, in certain cases,
rating products on measures defined by the researchers is not always a meaningful
activity for the user, for example when the user does not consider a quality dimen-
sion as relevant for the specific product or is unable to situate the question in her
own context.
We introduced personal attribute judgments and respective attribute elicitation
techniques as an alternative to standardized psychometric scales. These approaches
have the potential to account for the richness of user experience and bridge the
gap between formative and summative evaluation by quantifying rich qualitative
insight. We highlighted one technique, namely the Repertory Grid Technique which
has recently become popular in the field of Human-Computer Interaction and User
Experience.
We noted the lack of appropriate statistical analysis techniques that lead to prac-
tices that either treat RGT as yet another qualitative technique, or employ averaging
models which undermine the core motivation of RGT and relevant techniques. We
argued that the true value of RGT is in quantifying rich qualitative insights.
We identified the two dominant approaches in analyzing repertory grid data: a) a
qualitative, idiosyncratic approach with a narrative summarization, and b) a quan-
titative approach that employs averaging procedures using exploratory multivariate
techniques. We argued that these two approaches are limited in two respects. Quali-
tative approaches do not take properly into account the elicited attribute ratings and
 
Search WWH ::




Custom Search