Using Individual Observations to Estimate Recreation Demand Functions: A Caution