This paper considers the ways in which information can impact contingent valuation estimates of environmental values. The Hoehn/Randall distinction between value formulation and value statement is employed. We distinguish several ways information should be expected to effect true Willingness to Pay, and survey past contingent valuation work in the area. A model of information impacts is used to set up a series of testable hypotheses. These are then tested using data from a contingent valuation survey of the benefits of heathland preservation.