This paper models the recreation demand for Iowa wetlands, combining survey data on both actual usage patterns (i.e., revealed preferences) and anticipated changes to those patterns under hypothetical increases in trip costs (i.e., stated preferences). We formulate and test specific hypotheses concerning potential sources of bias in each data type. We consistently reject consistency between the two data sources, both in terms of implied wetland values and underlying preference parameters. Careful attention is paid to the interpretations of the test results, noting particularly how the interpretation of the same results can vary with the "school of thought" of the reader.