THE STUDY INCLUDED A QUESTIONNAIRE SURVEY OF ALL 45 PAROLE AGENTS SERVING ADULT CLIENTS IN IOWA, INTERVIEWS WITH A SUBSAMPLE OF 20 AGENTS, OBSERVATIONS DURING EXTENDED SITE VISITS TO 2 PAROLE OFFICES, AND A REVIEW OF 100 REVOCATION REPORTS. FOUR PROPOSITIONS WERE EXPLORED: THAT, AS THE PAROLE AGENT PERCEIVES OTHERS TO BE MORE TOLERANT OF PAROLEE DEVIANCE, HE OR SHE WILL BE LESS LIKELY TO DEFINE THE PAROLEE AS A DEVIANT; THAT PAROLE AGENTS MORE ORIENTED TOWARD PAROLE AS REHABILITATION (AS OPPOSED TO CUSTODY) WILL BE LESS LIKELY TO DEFINE PAROLEES AS REVOCATION CANDIDATES; THAT AGENTS WHO TEND TO VIEW PAROLEE INCIDENTS AS MORE SERIOUS WILL BE MORE LIKELY TO DEFINE PAROLEES AS REVOCATION CANDIDATES; AND THAT, AS THE AGENT DEVELOPS A MORE FAVORABLE IMAGE OF THE PAROLEE, HE OR SHE WILL BE LESS LIKELY TO DEFINE THE PAROLEE AS A REVOCATION CANDIDATE. A MAJOR FINDING IS THAT THE PAROLEE'S BEHAVIOR AND ASSESSED POTENTIAL PLAY A RELATIVELY MINOR ROLE IN REVOCATION DECISIONMAKING. AGENTS' REVOCATION DEFINITIONS WERE ONLY SLIGHTLY RELATED TO THE SERIOUSNESS OF PAROLEE MISBEHAVIOR. SUCH DEFINITIONS WERE INFLUENCED MORE BY REFERENT OTHERS (SUPERVISORS AND FELLOW AGENTS), ORGANIZATIONAL POLICIES, AND THE AGENT'S OWN ORIENTATION TOWARD REHABILITATION. IMPLICATIONS OF THESE AND OTHER FINDINGS RELATIVE TO THE NATURE OF THE REVOCATION DECISIONMAKING PROCESS, THE ROLE OF THE AGENT AS 'GATEKEEPER' IN THAT PROCESS, THE DECEPTIVE MEANING OF REVOCATION RATES, AND LABELING THEORY ARE DISCUSSED. SUPPORTING DATA, STUDY INSTRUMENTS, AND A BIBLIOGRAPHY AR INCLUDED. (LKM)