A method of paired comparison is adapted for use in estimating economic measures of value. The method elicits multiple binary choices for paired items in a choice set. Probability distributions and economic values are estimated nonparametrically and parametri-cally. The method is applied in an experimental context with a choice set composed of four private goods and several sums of money. The sample's median value estimates for the goods are generally not different than the market prices for these goods. People who are in the market for a good value it higher than those not in the market for the good.
