Recently, several problems with elicited utility functions have emerged. This paper concerns a fundamental problem in risk preference classification with elicited data. For the sample in this research, different functional forms resulted in reversals in preference classifications. This paper suggests that preference classifications must be interpreted with caution.