Using stated preference methods for valuation of non-market goods is known to be

vulnerable to a range of biases. Some authors claim that these so-called anomalies in effect render the methods useless for the purpose. However, the Discovered Preference Hypothesis, as put forth by Plott [31], offers an nterpretation and explanation of biases which entails that the stated preference methods need not to be completely written off. In this paper we conduct a test for the validity and relevance of the DPH interpretation of biases. In a choice experiment concerning preferences for protection of Danish nature areas from new motorway development, we find that respondent preferences are susceptible to starting point bias. In particular, our results show that the bias is gender-specific as only female respondents are significantly biased. Importantly, we find that the impact of the starting point bias decays as respondents evaluate more and more choice sets. This finding supports the Discovered Preference Hypothesis interpretation and explanation of starting point bias.