USING A SAMPLE OF 1,165 INVESTIGATIONS INTO POLICE CORRUPTION (AND OTHER FORMS OF SERIOUS POLICE WRONGDOING) WHICH WERE CONDUCTED DURING 1972 BY THE INTERNAL AFFAIRS DIVISION OF THE NEW YORK CITY POLICE DEPARTMENT, THIS RESEARCH EXAMINED WHAT ONE POLICE AGENCY HAS DONE ABOUT INVESTIGATING POLICE CORRUPTION. TO ACCOMPLISH THIS, THE INTERNAL INVESTIGATION PROCESS WAS STUDIED AS AN 'OPEN SYSTEM,' THAT IS, ONE WHICH RECEIVES AND PROCESSES INFORMATION OR COMPLAINTS OF POLICE CORRUPTION. IN ORDER TO EXAMINE MORE MEANINGFULLY THE INTERNAL INVESTIGATION SYSTEM, A NUMBER OF TAXONOMIC METHODS WERE EMPLOYED TO DERIVE EMPIRICAL TYPES OF POLICE OFFENSES. THESE TYPES WERE THEN USED AS INPUTS TO A MODEL OF THE INVESTIGATION PROCESS TO DETERMINE WHAT EXTENT THE SYSTEM'S OUTPUT (THAT IS, WHAT HAPPENED TO CASES WHICH WERE INVESTIGATED) WAS A FUNCTION OF DECISIONS CONCERNING HOW INVESTIGATIONS WERE CONDUCTED, TACTICS EMPLOYED, OR INFORMATION IN THE COMPLAINTS WHICH HAD SET THE PROCESS INTO MOTION. WHEN EXAMINING THE SYSTEM IN THIS MANNER, IT WAS FOUND THAT ALTHOUGH THE COMPLAINT DATA WERE REMOTE FROM THE CASE OUTCOMES--OFTEN EXERTING ONLY INDIRECT EFFECTS ON THE RESULTS OBTAINED--THE SYSTEM'S OUTPUT WAS CLEARLY AFFECTED BY THE INFORMATION PRESENTED IN COMPLAINTS. ALTHOUGH PROACTIVE INVESTIGATIVE METHODS ALLOW THE DEPARTMENT SOME CONTROL OVER INPUTS AND, HENCE OUTPUTS, AN ANALYSIS OF THE SAMPLE OF PROACTIVE CASES REVEALED THAT SELF-INITIATED INVESTIGATIONS DID NOT UNCOVER SUBSTANTIALLY DIFFERENT FORMS OF POLICE MISBEHAVIOR THAN THOSE BROUGHT IN COMPLAINTS. PROACTION DID, HOWEVER, RESULT IN BRINGING CRIMINAL OR DISCIPLINARY ACTIONS AGAINST OFFICERS AT OVER TWICE THE RATE OBSERVED FOR COMPLAINT-INITIATED INVESTIGATIONS. (AUTHOR ABSTRACT)