This study evaluates the effectiveness of a radon risk communication program based on how the estimated value of additional information varies across the six types of descriptive materials randomly assigned to a panel of homeowners participating in a radon utility model estimated with probit from respondents' answers to a contingent behavior question asking if they would purchase at a one-time price the services of a licensed technician to analyze their radon problems. The findings indicate that the information materials used most frequently by states and testing companies to explain radon's risk are the least effective of the six considered. Copyright 1990 by MIT Press.