A matter of time:the influence of context-based timing on compliance with well-being triggers

As occupational computer use grows, so do associated risks of negative health consequences. Repetitive strain injury and sedentary behavior are prevalent among computer users and both are associated with serious health risks. These risks can be minimized by healthy working behavior. One way to decrease the health risks associated with both of these conditions is taking short, frequent breaks during work, which are called microbreaks. Encouraging such behavior with persuasive technology (PT) has several advantages. Among the advantages are the availability and scalability of such interventions. Literature suggests that influencing people at the right time is critical to ensure the effectiveness of PT. Additionally, researchers have claimed that if a technology is aware of context, it should be able to identify such opportune moments and thereby increase compliance with the target behavior. An opportune moment to persuade, according to the Fogg Behavior Model, is one where the subject’s motivation and ability to perform the target behavior are at a high level. As such, a technology that tries to persuade knowledge workers to take microbreaks, should be able to determine when the motivation and ability of those workers to take a microbreak are at a high level. However, there appears to be a lack of empirical evidence for the Fogg Behavior Model, as well as for the claims of the importance of timing for PT and the usefulness of context information in determining opportune moments. The current research presents two studies. The first study was performed to assess whether context information (the computer activity of knowledge workers, such as the number of mouse clicks and key presses) can be used to make inferences about the level of motivation and ability to take a microbreak (e.g., overall computer activity is high, therefore the worker is too busy to take a microbreak). Six knowledge workers rated their level of motivation and ability to take a microbreak at different points in time, over the course of seven working days. Simultaneously, their computer activity was recorded. Confirming our expectations, the results show that moments of high and low (perceived) ability to take a microbreak can be partially predicted based on computer activity. More specifically, it can be based on two factors: the time since their last break and the change in their overall computer activity level. The level of motivation, on the other hand, could not be predicted based on computer activity. Next, the second study assessed whether presenting persuasive triggers at times of high ability leads to higher compliance, compared to times of low ability. A within-subjects experiment with 35 knowledge workers was carried out over the course of five working days. Each participant was subject to two conditions: triggers presented at times of high estimated ability and low estimated ability. The conditions were interactive and based on the current computer activity of the participants. Specifically, they were based on the two contextual factors described above.