Using choice experiment data for economic valuation we analyse how disbelief in survey information could affect the retrieved welfare estimates. We distinguish between two types of survey information to the respondents. The first type of information concerns the current environmental status of a water body. This information is provided prior to the valuation questions and the corresponding beliefs in the provided information are also elicited before valuation. The second type of information concerns the proposed improvements in the environmental status of the water body. We find that average welfare measures differ considerably according to whether respondents who disagree with the status quo levels and find proposed scenarios unlikely are included or not.