This study examines how errors in measuring marginal attribute pric es vary with the form of the hedonic price function. In simulations, consumers with known utility functions bid for houses with given attributes. Various forms of the hedonic function are estimated using equilibrium housing prices. Errors in estimating marginal attribute prices are calculated by comparing each consumer's equilibrium marginal bid vector with the gradient of the hedonic function. When all attributes are observed, linear and quadratic Box-Cox forms produce lowest mean percentage errors; however, when some attributes are unobserved or are replaced by proxies, linear and linear Box-Cox functions perform best.

Maureen L. Cropper, Leland Deck, Nalin Kishor, Kenneth E. McConnell, Valuing Product Attributes Using Single Market Data: A Comparison of Hedonic and Discrete Choice Approaches, The Review of Economics and Statistics, Vol. 75, No. 2 (May, 1993), pp. 225-232

We examine the effects of urban form and public transit supply on the commute mode choices and annual vehicle miles traveled (VMTs) of households living in 114 urban areas in 1990. The probability of driving to work is lower the higher are population centrality and rail miles supplied and the lower is road density. Population centrality, jobs-housing balance, city shape, and road density have a significant effect on annual household VMTs. Although individual elasticities are small absolute values (≤0.10), moving sample households from a city with the characteristics of Atlanta to a city with the characteristics of Boston reduces annual VMTs by 25%.