Appropriating Heuristic Evaluation for Mobile Computing - Human-Computer Interaction and Innovation in Handheld, Mobile and Wearable Technologies - page 32

Figure 4. Participants' demographics. Each value is ranked on a scale between 1 (min) and 4 (max)

Figure 5. Actual number of flaws, severity and distribution

Figure 6 can be used for further analysis of
how the specific heuristics from both sets fare
with regard to average severity and average num-
ber of flaws. Figure 6 indicates that mobile heu-
ristics are more effective in supporting the detec-
tion of flaws, while Nielsen's heuristics seem
better suited to cover the case in which high se-
verity flaws are present; also, mobile heuristics
seem to support a more detailed evaluation of the
mobile application (without considering the flaws
classified as catastrophic). It is worth noting that
some of the foregoing observations from Figure
6 are similar to those from Figure 5.
So far it might be observed that the mobile
heuristics produce a more accurate evaluation in
terms of number of problems detected (more flaws

are identified), reduced variation among experts'
analyses, and problems' severity ranking (this is
actually also supported by the qualitative data
collected during the evaluation, where most ex-
perts said that Appl.1 was much better designed
for a mobile use when compared to Appl.2). Thus
the mobile heuristics tend to focus the evaluation
on the mobile issues instead of directing experts'
attention at a more general level (although the
kind of setting we used in this study promoted an
evaluation of applications that was more func-
tionalities-based than contextual). Moreover, the
mobile heuristics could be applied when/where
the extreme flaws have been addressed or are not
an issue in the design. If such flaws have to be
identified before proceeding, mobile heuristics