In order to explore why the most widely used measure of risk proneness (Choice-dilemmas Questionnaire) yields so much internal variation, its items were filled out in the usual manner and then in terms of how serious the consequences of failure would be in each case. Correlations between the usual risk recommendations, consequences of failure, and national norms for magnitude of risky shift suggested a positive relationship between risk preference and degree of shift toward risk, a negative relationship between seriousness of consequences of failure and risk preference, and a negative relationship between seriousness of consequences and degree of shift toward risk. The findings suggest empirical and theoretical specificity for the speculation that different items engage different values.