Conclusions - Modeling Users' Experiences with Interactive Systems - page 139

such approaches. First, they rely on an assumption that the relevant latent constructs
remain constant, but their perceived value and relative dominance change over time.
As it was shown in chapter 4, however, prolonged use might relate to a different
set of user experiences such as daily rituals and personalization that do not become
apparent in users' initial interactions and that may thus be not captured by the mea-
surement model. Reversely, other constructs that are salient in initial use situations
may become irrelevant over prolonged use. This may lead to distorted data, for in-
stance when participants are not able to interpret the personal relevance of a given
measure (e.g. learnability) in the current usage situation. Last, such approaches pro-
vide rather limited insight into the exact reasons for changes in users' experiences.
They may, for instance, reveal a shift in the dominance of perceived ease-of-use and
perceived usefulness on intention to use a product (e.g. Venkatesh and Davis, 2000),
but provide limited insight to the exact experiences that contributed to such changes.
Insights into these rich and contextualized data is what designers need in designing
for a given context.
We argued for an alternative approach to the measurement of the dynamics of
experience over time, that relies on a) the elicitation of idiosyncratic self-reports of
one's experiences with a product, the so-called experience narratives , and b) the
extraction of generalized knowledge from a pool of experience narratives through
content analytical procedures. In this sense, each narrative may provide rich insights
into a given experience and the context in which it takes place. However, generalized
knowledge may also be gained from these experience narratives. Such generalized
knowledge may be reflected in questions like: how frequent is a certain kind of
experience, what is the ratio of positive versus negative experiences and how does
this compare to competitive products, how does the dominance of different product
qualities fluctuate over time and what should we improve to motivate prolonged
use?
This leads to two research problems. Firstly, how can we elicit experience nar-
ratives efficiently? Chapter 5 reviewed existing methodological paradigms for in-
quiring into users' experiences over time. Longitudinal designs such as the one em-
ployed in chapter 4 were identified as the gold standard in the study of long-term
effects of product use, but their labor-intensive nature was highlighted as a barrier
towards their adoption but also as an antecedent of restricted samples in terms of the
user population, the product population and studied time. An alternative method-
ological approach was proposed that relies on the elicitation of user' experiences
with a product from memory. The chapter presented iScale, a tool that was designed
with the aim of increasing participants' effectiveness in recalling their experiences
with a product. Two different versions of iScale, the Constructive and the Va l u e -
Account were motivated by two distinct theoretical approaches in the reconstruction
of one's emotional experiences. These two versions were tested in two separate
studies against a control condition, free-recall employing no process of enhancing
users' recall. Overall, iScale was found to result in a) an increase in the number of
experience reports that participants provided, b) an increase in the amount of con-
textual information for the reported experiences, and c) an increase in participants'
accuracy in recalling concrete details of the experienced events, thus suggesting that