There have been inconsistencies in past residential water demand studies concerning the specification of the relevant price variable. Both average and marginal price have been used. Households faced with the decision to expand (or contract) their water consumption are responsive to the additional cost (or savings) involved with this decision. The appropriate measure of this price is marginal price, not average price. Data were gathered from a large metropolitan area, Miami, Florida, to illustrate the divergence in estimation from two demand models, identically specified, except for the price variable. One utilized marginal price, the other average price. Predictions of consumption varied from 22 to 107% between the two models. The average price model significantly overestimated the response of consumption to price and income changes, a result worthy of consideration.