Biases in contingent valuation exercises are often difficult to quantify and to correct due to the hypothetical nature of the method itself. One common problem is hypothetical bias, which in recent years has been addressed in various ways, particularly through the inclusion of a certainty question in the questionnaire. We analyse the impact of applying differing corrections based on the certainty question for double-bounded dichotomous economic valuation exercises. The empirical application is based on data gathered from four surveys conducted with various groups to obtain the value allocated to a modern art museum: the Museo Patio Herreriano de Arte Contemporáneo Español, in the city of Valladolid (Spain). Findings indicate that estimates of Willingness to Pay are significantly reduced when a higher certainty threshold is required, in other words when individuals are more certain of their choices. Furthermore, correction through recoding proved more severe than correction through exclusion.