Information Technology Reference
In-Depth Information
5.4.1.5
Dependent Variables and Expectations
Despite the fact that the study was explorative in nature, a number of predictions
about the differences in performance of the three versions of the tool can be made.
Number of elicited experience reports
Based on existing evidence that the reconstruction of events in a serial chronological
order cues the recall of temporally surrounding experiences and related contextual
cues (Anderson and Conway, 1993), it was expected that the constructive iScale will
result in an increase in the number of experiences being reported. For the value-
account iScale, which makes it more difficult for participants to reconstruct their
experiences in a chronological order, the difference to the control condition was
expected to be smaller.
Richness of elicited experience reports
Similar to number of elicited experience reports, we expected that reconstructing
in a chronological order would lead to more contextual cues, thus providing richer
insight into users experiences. Such contextual information may relate to temporal
(i.e. when did the event happen), factual (i.e. what happened), social (i.e. who was
present) and others. To identify these different factors of richness, we submitted the
experience reports to a qualitative content analysis (Hsieh and Shannon, 2005) (see
section 5.4.2.2 for a more elaborate description of this process).
Test-retest consistency in time estimation
As participants are expected to recall more contextual cues in the constructive iS-
cale, this should increase the test-retest consistency in recalling factual details of the
past experiences, such as temporal information (e.g. when did the experience take
place) (Kahneman et al., 2004). We further predict that graphing in general (even in
the value-account condition) will result in a more consistent recall of such temporal
information, as graphing provides a temporal overview of the recalled experiences.
To assess this, we coupled experience reports from the two sessions of the study
that referred to the same experience and computed the difference in estimated time
across the coupled experience reports (see section 5.4.2.3 for a more elaborate de-
scription of this process).
Test-retest consistency of graphs
The test-retest consistency of the participants' graphs (i.e. value-charged informa-
tion) was expected to be higher in the value-account version, where participants cue
this directly through a hypothetical memory structure, compared to the construc-
tive version, where participants are assumed to reconstruct this information from
concrete contextual cues recalled from episodic memory. This is based on the as-
sumption that repeated chronological reconstruction might cue a different set of
experiences and, thus, lead to a different path in reconstructing the overall pattern
Search WWH ::
Custom Search