The randomized response approach is designed to elicit truthful responses by protecting the anonymity of the subjects in their responses to sensitive questions. The present study surveyed 503 undergraduate students in sociology courses at an Ivy League university. The survey asked how many times the respondents had committed 20 offenses, including vandalism, drug use, rape, arson, and robbery. A total of 161 students completed the questionnaire directly in group sessions. The other 337 students answered the same questions using a randomizing technique entailing the use of two dice, one green and one red. The respondents were instructed to throw the dice prior to each question. If the green die showed 1, 2, 3, or 4, the question was to be answered truthfully, If the green die showed a 5 or 6, the subject was to look at the red die and record the number that appeared. If the red die showed 6, the respondent was instructed to write 0. The generally similar results for the two groups could mean that the group administration of the direct questionnaire resulted in little perceived threat to anonymity. In addition, respondents may not have perceived the questions to be sufficiently sensitive to provoke deliberate falsification. Finally, the apparently low level of offenses among the study population may have prevented the identification of differences between the two groups. Further research should focus on these issues. Data tables, a list of the offenses on the questionnaire, and 32 references are provided.