The study tests two behavioral intervention programs that were aimed at changing energy-related behaviors in an office setting. Participants were 93 office employees in three different departments of a construction company. Each department was randomly assigned to a four-week intervention to one of the following three conditions: control, intervention program or intervention program with group identity salience. The first intervention condition consisted of goal-setting, feedback, information and prompts. The second condition was identical to the first, with an additional group-identity manipulation. Three different types of data were collected: consumption of electricity and paper, pre- and post-questionnaires and weekly observations. The results show that the employees in the two experimental conditions perceived themselves to have changed their behaviors more than the control group. This is supported by the observational data for the intervention program, but could not be shown in the measures of energy consumption, which showed a decrease for all three departments. The implications of these results are discussed. (c) 2015 Elsevier Ltd. All rights reserved.