In some recent contributions to the literature on public choice both in this journal and elsewhere, Lovell (1975, 1978) has demonstrated that if it is presumed that the median voter dominates the ballot box and if it is assumed that all voter have a particular identical, but very generalized, utility function, then a particular empirical specification for explaining the proportion of income spent on schooling across school districts is implied. Moreover, this specific empirical form is capable of yielding estimates both of elasticities of substitution between schooling and all other goods, as well as estimates of the degree to which non-residential taxes are shifted to non-voters and the degree to which renters do not perceive the property tax burden. The estimate of the elasticity of substitution has important implications for certain changes in the institutions of public choice. The purpose of this note is to repeat, to the extent that it is possible, Lovell's empirical experiments on a separate data base and to reiterate the implications of our findings, which are consistent with his, that the elasticity of substitution between schooling and all other goods is less than unity.