The endowment effect, which predicts undertrading and a willingness-to-accept greater than willingness-to-pay, is studied using responses that remove all reference to buying or selling and focuses only on choice tasks. The results significantly lower the willingness-to-pay/willingness-to-accept discrepancy, but the latter is still significant. A high efficiency open display uniform price auction is used to exchange mugs for money. Since mugs are randomly assigned to half of 2N subjects, N/2 mugs are predicted to trade. Less than N/2 mugs trade on average, but more than previously reported. The phenomenon exists but is less prominent than reported previously.