Fiddling with Thermostats: Energy Implications of Heating and Cooling Set Point Behavior

Household energy use depends heavily on how people set their thermostats. In fact, evaluations of the benefits of new cooling and heating technologies often assume specific thermostat behaviors, or set points. California's Title 24 Standards, for example, assume a certain range of settings and frequency of daily changes in those settings. Until recently, data have not been available to test such assumptions. In 2001-02, the California Energy Commission conducted a demand response experiment that produced unique, high frequency observations of residential thermostat settings and internal temperature measurements, which allow testing of assumptions about thermostat behaviors. Comparing the thermostat settings observed in the California experiment with those commonly assumed in policy modeling indicates that people change cooling and heating set points much more frequently than has been assumed. Frequent set point changes, and the extreme diversity of set point behavior across the population, have significant energy implications. This paper uses Shannon Entropy to assess consistency of thermostat settings, which can produce both higher and lower levels of energy consumption than is conventionally assumed. The findings call into question the benefits of energy efficiency programs that focus on equipment replacement and choice.