Sample Selection Bias in Estimating Housing Sales Prices

This paper explores the bias in estimates of housing price appreciation that arises because of problems of sample selection. As suggested by Haurin and Hendershott, sample selection potentially is a serious problem because studies normally are based on samples of only homes that have sold, not all houses. Using the procedure developed by Heckman, the empirical results of this study provide confirmation of the significance of the housing sample selection problem.