Information Technology Reference
In-Depth Information
Specifically, additional information gained from the Q Sensor for this study
redirected and clarified the impressions that we had based on observation and
the participants' thinking aloud. Based on those measures, we expected that the
digital textbook would have been more stimulating based on how the partici-
pants described and interacted with it. However, Q Sensor data revealed that
participants had a higher arousal level while using the paper textbook. Other
qualitative data indicated a negative direction for those emotions, as partici-
pants struggled with their tasks. This negative emotion was stronger than the
pleasurable emotions felt while using the digital textbook.
This approach would be ideal for projects whose goal is to understand partic-
ipant emotional responses and the severity of those reactions throughout their
interaction with a product. By associating metrics across data sets, researchers
can pinpoint a participant's exact emotional reaction, and what was causing that
reaction, at any point during their session. This unique combination of met-
rics provides a new window into a participant's emotional reactions above and
beyond what is articulated during a standard think-aloud usability study.
However, these techniques require additional time to set up the study appro-
priately and to analyze the results. For example, researchers will want to plan on
using time markers with the Q Sensor. We learned during our data analysis that
we could have saved significant efforts postanalysis by adding more markers to
Q Sensor data during the sessions. For this study, the team spent approximately
2 work weeks scrubbing, combining, and analyzing the data. However, a more
recent project that's used these same techniques only took us 3 work days as we
used more Q Sensor markers during the sessions. This method probably won't
make sense for a basic formative usability study, but we believe would offer ben-
efits for projects with a larger scope.
We are continuing to refine and build out these techniques through addi-
tional projects and are applying them to new domains.
ACKNOWLEDGMENTS
Thanks to the Design and Usability Center at Bentley University for their support,
to Affectiva for use of the Q Sensor and analysis support, and Pearson Education
for providing the digital and printed textbooks. Also, thanks to Vignesh Krubai,
Diego Mendes, and Lydia Sankey for their contributions to this research.
REFERENCES
Search WWH ::
Custom Search