Using the Q-type correlation coefficient in conjunction with the Q-sort relative rating procedure, the specific issues to which this investigation addressed itself were (a) whether staff members would exhibit high agreement, (b) whether inmates would exhibit high agreement, and, most important, (c) whether staff and inmates would exhibit low agreement in their appraisals of the prison's official administrative code regulating inmate conduct. These expectations were derived directly from the opposition assumption. Results substantiated the first two expectations. The third and most important expectation, however, was convincingly disconfirmed. Agreement between the staff groups and the inmate groups was as high as the within-staff and within-inmate agreement. Moreover, some groups of staff showed more agreement with some groups of inmates than they did with other groups of staff. Likewise, some groups of inmates showed more agreement with some groups of staff than they did with other groups of inmates.(Publisher abstract)