In August 1995 a 50-item questionnaire was mailed to a randomly selected sample of 400 residents of Hamilton County (Cincinnati), Ohio. The study used Dillman's (1978) "total design method" with slight modifications. A total of 237 usable questionnaires were returned. Taken together, the results suggest that the demographic differences between the 1986 and 1995 Cincinnati samples are unlikely to account for attitudinal shifts that may have occurred over time. The survey queried respondents on their view of the main purpose of prison, support for treatment programs and their expansion, support for various types of treatment programs, and views of various correctional policies. The data show that public support for rehabilitation has noticeably declined over the past decade. Since 1986 support for rehabilitation as the main goal of imprisonment decreased by more than 20 percentage points; whereas, support for punishment has increased by this same amount. In addition, the public has become less supportive of specific modes of rehabilitation, such as psychological counseling and educational and vocational training; it strongly opposed the use of early release. Despite this declining public support, however, the rehabilitative ideal continued to show considerable tenacity. A third of the American public still believed that rehabilitation should be the main emphasis of prisons; more than half of the citizens endorsed expanding programs; 40 percent supported early release for good behavior and participation in treatment programs; almost 70 percent favored work with restitution; a large proportion remained optimistic about the possibility of rehabilitating juveniles and nonviolent offenders; and a majority still viewed treatment programs as the best policy for dealing with offenders while they are incarcerated. Apparently, however, the rehabilitative ideal has been eroded by two decades of the sustained attack of "get tough" rhetoric. 4 tables, 2 notes, and 47 references