The study found that “working alliance” scores ranged from 110 to 186 (maximum of 252), with a mean of 160.4. The definition of “working alliance” used in this study was developed by Bordin (1979). Bordin views the working alliance as consisting of three essential components: the level of agreement on therapeutic goals; the ability to engage collaboratively in mutually negotiated tasks intended to meet these goals; and the establishment of a trusting, mutually respectful relationship that provides the client with a safe context for exploring personal difficulties. Two of the predisposing variables--services to men and level of education for the clinician--were significantly related to perception of working alliance. Clinicians who provided treatment to women tended to have a stronger working alliance than clinicians who treated men also, clinicians who did not have master’s degrees tended to have a greater working alliance with inmates than clinicians with master’s degrees. Region of the State was also significantly related to clinicians’ perception of working alliance. Region may be serving as a proxy for some other variable, such as race. This should be explored in future research. Data were obtained from staff members at 17 prisons where mental health services were provided to inmates (n=107). A questionnaire was developed for the assessment of clinicians’ perception of the overall working alliance with inmates in their caseloads. Various clinician and client variables were measured and compared with perception of working alliance. 5 tables and 40 references