Persuasive computing systems have received an outburst of interest both in research and practice over recent years, with a growing body of literature attempting to understand how people use these devices to collect, monitor, reflect, and take action upon, their behaviors. This is accompanied by the increasing prevalence of these devices in the everyday life, across diverse goals. Persuasive systems are used to understand routines, learn something out of interest or simply have fun and satisfy curiosity – extending the conventional vision of mediators of change.
Within our own work, we have contributed to a better understanding of how persuasive systems are used in everyday life and how, in turn, they shape individuals’ behaviors. Our research has been guided by two, intertwined goals:
First, we have investigated the real life practices emerging from owning and using persuasive systems – i.e., how these devices fit into the everyday life and routines of people that use them.
In a fist attempt, we developed Habito, a mobile activity tracking prototype. Habito logged users’ activity and interactions with the app (e.g., the frequency and nature of usage sessions, as well as the frequency and distance covered during walking activities). We conducted a 10-month study of the adoption, usage and discontinuation of Habito by a sample of 256 users. The data allowed us to gain quantitatively informed insights into the adoption rates of these devices, as well as their daily use.
Our initial insights were complemented with in-situ observations of the daily use of these devices. In a recent study, we used wearable cameras to record the daily use of activity trackers. We collected and analyzed 244 incidents where activity trackers were used.
This data was combined with participants’ physical activity data, providing insights into how tracker use is enmeshed within everyday life. Most notably, we found activity tracking, as a diversified practice, dependent upon and threaded into what goes on around us.
Second, we have investigated the effects of theoretically informed design strategies for behavior change, such as deliberate goal-setting, and the ways in which those effects vary with individual differences, design choices, and contexts of use.
To this end, we have adopted a micro level of analysis, investigating short-term outcomes of the use of persuasive systems. We have tracked behavioral outcomes – such as users’ informational needs, and physical activity, as they interact with specific components of these systems. This approach has allowed us to gain insights as to when a component has an effect on behavior, and how that effect changes over time.
For instance, in our research on glanceable displays, we developed a smartwatch interface (Normly) which provided normative feedback on goal completion – namely, by comparing one’s goal completion to that of others with a similar goal. In a field study, next to monitoring physical activity, application usage was logged, including when our interface was checked, as well as interactions within.
This data allowed us to draw insights regarding the impact of our interface. For instance, we found users’ informational needs and interactions to evolve in regards to their performance. Users were found to be more likely to initiate a new walk when closely ahead or behind of others. This likeliness dropped as these distances grew.
Activity Tracking in vivo
Rúben Gouveia, Evangelos Karapanos, Marc Hassenzahl.
CHI, 2018. pdf