| The lifestyle people want to lead is often different than the lifestyle they do lead. People want to be financially secure, yet consumer debt is on the rise. People want to be fit and healthy, yet physical inactivity and poor eating habits are leading to serious health problems. A recent trend in computing is to use technology to help people change their everyday behaviors in support of the lifestyle they want to lead. The focus of this dissertation research is to develop and investigate a technology to support lifestyle behavior change. This dissertation introduces a computing system that has been designed to encourage individuals to participate in regular and varied physical activity. The computing system, UbiFit, was designed based on theory as well as findings from recent empirical investigations that have used technology to encourage healthy behavior change. The UbiFit system uses on-body sensing, real-time activity inference, an electronically assisted mobile journal, and a novel personal, mobile display to encourage physical activity. The iterative design and evaluation of the UbiFit system is described and results from the three-phased evaluation, which included a paper-based survey (n=75), 3-week field trial (n=12), and 3-month field experiment (n=28), are presented.;This dissertation makes several contributions. For example, results reveal that the metrics that are traditionally used to evaluate activity inference systems do not sufficiently capture the user's perspective. This dissertation describes the types of inference errors that users perceived and how those errors affected the system's credibility. It also shows that allowing users to add to, edit, and delete inferred data helps to improve the system's credibility, especially when inference errors are experienced. Based on the detailed activity logs that were collected during the two field studies, this dissertation raises the question of what it means for physical activity behavior to be "regular" or "consistent," as participants' weekly activity levels tended to be inconsistent when using a standard definition from the medical literature. Perhaps most exciting was that the personal, mobile display was effective at helping individuals maintain their physical activity level, even during the winter holiday season which is notorious for physical inactivity. |
