Field Test of Energy Information Feedback

Abstract

The Energy Information Administration estimates that, in 2007, U.S. domestic passenger vehicles burned 113 billion gallons of fuel and thus generated more than 16% of U.S. greenhouse gas emissions. Past field experiments and simulations suggest that energy information feedback to drivers could have spared 10% to 25% of those gallons. However, the theoretical underpinnings of past experiments have primarily been ad hoc, with application of their results limited to specific conditions of the experiment and feedback design. More rigorous behavioral theory would allow researchers to account for more variation in driver response to feedback, create testable hypotheses about the effectiveness of current systems, and provide a basis for designing more-effective systems. This paper presents drivers’ responses to energy feedback in a field test involving 98 participants from 43 households in California and compares the results with the concepts that underlie the theory of planned behavior and the extended model of goal-directed behavior. About 40% of participants reported more economical driving behaviors after viewing the feedback; estimation of actual changes in fuel use is left for future research. After viewing real-time energy information, numerous drivers reported setting goals, having emotional reactions, and creating new driving behaviors. Distraction from the primary driving task was a persistent problem for some drivers. Web-accessible information was not as motivating to participants. Finally, the study finds evidence of correspondence between theoretical behavioral factors and drivers’ responses.

The impact of numerical vs. symbolic eco-driving feedback on fuel consumption – A randomized control field trial

“…Consequently, the reliability and validity of many study outcomes may be challenged. Along with a handful of other examples (e.g., af Wahlberg, 2007;Stillwater and Kurani, 2011;Tulusan et al, 2012), we consider our experiment to be one of the few studies that investigates feedback on eco-driving under realistic driving conditions, with a larger sample size, a research design that is strong in causal inference, and an explicit reflection of important covariates (road attributes). Therefore, we call for future research to adopt our approach and even further increase the sample size to ensure validity of the research outcomes.…”