Using animated social feedback to motivate air conditioning energy saving

In 2009, Midden & Ham found that social feedbacks motivated users to reduce energy consumption in washing experiments with a simulated washing machine. They also found that having negative social feedbacks reduced energy consumption better than having factual, positive social, and no feedbacks. The present study compares the effect of positive and negative social and factual feedbacks on user preferences in air conditioning (AC) temperature setting. A higher room temperature setting means a lower energy use. We predicted that having any of the feedbacks increased chosen temperatures compared to having no feedback and that having negative social feedbacks increased chosen temperatures compared to having either factual, positive social, or no feedbacks. An interface was made in order for the AC remote controller to be able to communicate with computers. Animated faces which are able to smile or frown were made representing positive and negative social feedbacks. The factual feedback is represented by a bar indicator that turns red when users choose low temperatures and green when high. 159 participants were asked to participate for the experiment and randomly assigned to either control, factual, positive, or negative feedback groups. A computer screen was put in front of respondents to display the animated feedbacks. Participants set their preferred AC temperature by means of a remote controller and the screen deliver feedbacks according to the chosen temperature. One-way ANOVA were used to analyze the data with temperatures as the dependent variable and feedbacks as independent variables. Results show that there was a significant effect of feedbacks on chosen temperatures, F(3, 153) = 20.96, p <; .001, r = .54. There was a significant linear trend, F(1, 153) = 60.55, p <; .001, r = .53, indicating that as the feedback changed from factual to positive and negative social, the temperature increased proportionally. Planned contrasts showed that having animated social feedbacks increased chosen temperatures compared to having no feedback, t(152) = 6.80, p <; .001 (1-tailed), r = .39, and that having negative social feedbacks increased chosen temperatures compared to having positive social feedbacks, t(152) = 2.14, p <; .05 (1-tailed), r = .17. However, having factual feedbacks did not increase chosen temperatures compared to having no feedback, t(152) = 1.10, p > .05 (1-tailed), r = .09.