Residential energy consumption: An analysis-of-variance study

Abstract

In this report, tests of statistical significance of five sets of variables with household energy consumption (at the point of end-use) are described. Five models, in sequence, were empirically estimated and tested for statistical significance by using the Residential Energy Consumption Survey of the US Department of Energy, Energy Information Administration. Each model incorporated additional information, embodied in a set of variables not previously specified in the energy demand system. The variable sets were generally labeled as economic variables, weather variables, household-structure variables, end-use variables, and housing-type variables. The tests of statistical significance showed each of the variable sets to be highly significant in explaining the overall variance in energy consumption. The findings imply that the contemporaneous interaction of different types of variables, and not just one exclusive set of variables, determines the level of household energy consumption.