iScale: Studying Long-Term Experiences through Memory - Modeling Users' Experiences with Interactive Systems - page 112

Figure 5.10 depicts the relationship between the reported (actual) time versus the
perceived time (i.e. the position of the node along iScale's x-axis). To enable an
across-subject comparison, we normalized the reported (actual) time variable by the
total time of ownership of the product for each participant, resulting to an index from
0 to 1. Given no accessibility bias, one would expect a linear relationship between
these two pieces of information. One might note in figure 5.10a, however, that the
variance in the dependent variable (actual time) is not uniformly distributed across
the range of the independent variable (position along the x-axis of iScale). If one
transforms the variables by the logarithmic function, the data become much more
uniformly distributed. A linear regression on these transformed variables shows a
significant prediction accounting for 66% of the variance in the dependent variable.
This suggests a power law relation between the recalled actual time of the experi-
enced event and its position along the graphing tool's timeline with a power equal to
1/1.47=0.68 (i.e. perceived-time = actual-time 0 . 68 ). In other words, participants had
a tendency to use a substantial fraction of the x-axis of iScale to map their initial
experiences. In a similar vein, 95% of all experience reports related to the first six
months of use and 75% of all experience reports related to the first month of use
whereas the median time of ownership was 10 months. It thus becomes evident that
experiences pertaining to initial use are more accessible in participants' memory. To
account for this accessibility bias we compute the temporal distance between two
events through formula 5.1.