Personal Attribute Judgments - Modeling Users' Experiences with Interactive Systems - page 21

the participant to motivate its importance, e.g. “why is expressiveness important to
you?” Pyramiding, on the other hand, also known as negative laddering, seeks to
understand the lower level attributes that make up for a given quality; in pyramiding
the researcher asks the participant to elaborate on what makes the given product to
be characterized with the respective attribute, e.g. “what makes this product more
[easy to use]?”. This process can be repeated for all possible combinations of prod-
ucts and until no new attributes arise. The result is a list of attributes that the spe-
cific individual uses to differentiate between a set of products. The attributes may
then be employed in rating scales, typically Semantic Differentials (Osgood et al.,
1957), and each participant rates the set of products on her own elicited attributes.
Participants' ratings are subsequently analyzed with exploratory techniques such as
Principal Components Analysis (PCA) or Multi-Dimensional Scaling (MDS).
With the recently increased interest in user experience (e.g. Hassenzahl and
Tractinsky, 2006), the RGT has become popular in the field of HCI. Hassenzahl
and Wessler (2000) employed the RGT to evaluate the outcome of parallel design
and analyze the perceived character of websites (Hassenzahl and Trautmann, 2001).
Fallman and Waterworth (2005) elicited users' experiences with mobile technology
devices, while Davis and Carini (2004) explored player's experience of fun in video
games. Hertzum et al. (2007) studied the differences between designers' and users'
perceptions for three diverse cultural settings while others have used the RGT as an
ethnographic method in understanding how individuals organize their documents
and communications (e.g. Szostek et al., 2008; Bondarenko and Janssen, 2009). It,
thus, becomes evident that an increasing number of researchers in HCI, empha-
size the idiosyncratic nature of subjective judgments on the quality of interactive
products.

2.3 The Quantitative Side of Repertory Grid - Some Concerns

While the RGT has become popular in the field of Human-Computer Interaction
and User Experience, one may note several problems in the application of RGT in
practice, at least from a quantitative perspective. Below we highlight these issues
and introduce the contribution of our work in the analysis of Repertory Grid data.

2.3.1 Are We Really Interested in Idiosyncratic Views?

The RGT originates from clinical psychology where the emphasis is on an individ-
ual's perceptual and cognitive processes. In the field of user experience, however,
the interest is not in the idiosyncratic views of an individual but rather on some
more-or-less homogeneous groups of individuals. Due to this focus on idiosyncrasy
within the field of personal construct psychology, interpersonal analysis of repertory
grid data has received very little attention. We have come to realize two inadequacies
of the RGT when used for interpersonal analysis.
First, the interpersonal classification of attributes is often performed purely on se-
mantic grounds, i.e., without testing congruence on the rating scores. Such a practice