This report summarizes the econometric theory surrounding estimation of residential demand functions and describes the methodology used to develop residential electricity demand function for 136 utility service areas. A single equation function is described; its functional form is log-linear. Special treatment is given cost- of-living differences and climatic variables. The demand equations were estimated in two stages. The selected model contained humidity as a second weather variable, is adjusted for heteroskedasticity, and has all variables correct in sign and approximate magnitude. Elasticites are reported for income and prices and compared to previous studies. Price and income projections are then used to predict 1985 electricity sales.