Drawing on framing theory, the present study tests the impact of question wording on people's reported opinions about the harshness of their local courts. A randomized experimental design tested two salient variations against the standard wording used in the National Opinion Research Center's General Social Survey (GSS). The results indicated statistically significant differences, with fewer respondents expressing a desire for greater harshness with the alternative forms than the standard question form. Four of the five correlates that the authors examined also showed differential relationships with punitiveness among the question forms. These findings suggest that scholars should carefully consider the meaning of people's responses when interpreting the GSS question as an indicator of public punitiveness. (Published Abstract)