A total of 905 usable responses were received, for a 43-percent return rate. Analysis of data indicates that there was neither unanimity of opinion nor unidimensionality in the differences of opinion. Across levels of attitude, the correspondence was fairly strong: those supportive of AA at the general level also were supportive of AA at the specific level. In general, nonwhite and female respondents more strongly supported AA positions than did whites and males. The majority of respondents, regardless of race and sex, agreed on the acceptability of minority coworkers and to a lesser extent, of female coworkers. Even with sex and race controlled, correctional employees showed more supportive attitudes than did law enforcement employees. A number of correlations between attitudes and demographic and individual variables were found, particularly with respect to education. There was greater support for general and specific pro-AA dimensions among employees whose agencies were successful in meeting AA objectives, although these differences disappear when only respondents in agencies with enforced quotas are examined. Respondents in agencies with enforced quotas or meeting AA goals expressed less general job satisfaction, and less satisfaction with internal communication and the promotional system. Implications of findings for AA programs are discussed. Extensive tabular data are provided.