Factor analysis resulted in a 36-item Likert scale with the items correlating at least .47 with a general factor. The ATP scale possesses moderate to high split-half (r=.84 to .92) and test-retest (r=.82) reliability. No evidence of response distortion was found. A number of comparisons between selected groups provided considerable evidence of validity. Groups of prisoners, as well as persons engaged in prisoner rehabilitation or prison reform, scored highest on the scale. Intermediate-scoring groups consisted of students, a community sample, and correctional officers. As predicted, the group with lowest average ATP scores consisted of law enforcement officers. (Publisher abstract)