We develop a general norm-dependent utility function with disutility for actions that are inferior or superior to a norm. We test its validity by assessing the moderating role of norm sensitivity in explaining responses to peer influences in a budget reporting experiment. Managers become less honest after seeing a less honest peer (the rotten apple effect) and more honest after seeing a more honest peer (the sterling example effect). We measure the sensitivity to social norms by the Maintaining Norms Schema score generated from the responses to the Defining Issue Test-2 moral reasoning questionnaire. We find that (1) the sterling example effect is significantly increased in an individual’s sensitivity to social norms and (2) the rotten apple effect does not vary significantly with an individual’s sensitivity to social norms. Our evidence supports inclusion of a disutility component for actions that are inferior to the norm in representations of personal preferences.
JEL Classifications: C72, D03; J44; M41; M55.