Information Technology Reference
In-Depth Information
3.2
The Study
The data for the present analysis was taken from a study of Heidecker and Hassen-
zahl (2007) of individuals' perceptions of eight university websites (figure 3.1). The
study was part of a larger project aiming at understanding how the Technical Univer-
sity of Darmstadt (TUD) is perceived in comparison to other regional competitors.
Ten individuals, all students at TUD, participated in the study.
Ta b l e 3 . 1 Attribute categories and examples
Attribute category
Example
Layout
Graphical layout - Textual layout
Colorful - Pale colors
Professional - playful
University Image
Technical studies - Social studies
Emphasis on achievement - Average univ.
Refers to student life - Modern organization
Information Access
Fast access to information - time-intensive
Legible - Tangled
The eight university websites were presented to participants in the form of col-
ored A4 screenshots of the main page. Using the Repertory Grid Technique, a num-
ber of attributes on which the eight websites differ, were elicited from each par-
ticipant. Participants were then asked to rate the websites on their own elicited at-
tributes, using 7-point Semantic Differential scales. The resulting data set consisted
of a total of 118 attributes (10 to 14 per participant) on which ratings for the eight
different stimuli were elicited.
3.3
A Multi-dimensional Scaling Approach to Account for
Diversity
The starting point of the proposed approach is that of identifying the different views
that each participant has on the set of products. In this step, an average model is
attempted for each participant. However, attributes that are not adequately predicted
by the average model (see Table 3.2) are removed and used in deriving a second
model, i.e. a second view for the specific participant (Figure 3.2 illustrates two di-
verse views derived for one participant).
Once the diverse views of all individuals have been identified, the similarity
among them is assessed and views are clustered into groups of increased homo-
geneity.
A final set of diverse configurations is formed by grouping the similar views,
which are then used to model the attributes from all participants.
 
Search WWH ::




Custom Search