THE DATA BASE FOR THE FINDINGS CONSISTS OF INTERVIEWS MADE IN 1973 WITH MEMBERS OF POLITICAL ELITES IN THREE STATES -- FLORIDA, ILLINOIS, AND WASHINGTON. IT WAS INTENDED TO MEASURE THE PROPENSITY OF EACH STATE ELITE TO ACCEPT PRISON REFORMS. THE DATA INDICATE A DIVISION OF AUTHORITY SO THAT ACROSS STATES SOME GROUPS DEFINE CORRECTIONS AS THEIR RESPONSIBILITY WHILE OTHERS DO NOT. THE SAME GROUPS WERE NOT EQUALLY ACTIVE IN ALL STATES. STATES WERE SHOWN TO DIFFER IN THE DEGREE TO WHICH THE CONFIGURATION OF INFLUENCE WAS BASED ON PERSONAL CONTACT BETWEEN ACTORS. SOME STATES HAVE STRUCTURES IN WHICH A LARGE PROPORTION OF INTERESTED PEOPLE ARE CONTINUALLY IN TOUCH WITH ONE ANOTHER. OTHER STATES EITHER WERE LESS POLITICALLY INTEGRATED OR COORDINATED ACTIVITY THROUGH DIFFERENT TECHNIQUES. IN STATES WHERE THERE IS A NETWORK OF CLOSE ASSOCIATES IN A DOMINANT COALITION, REFORMS ARE ACCOMPLISHED MORE EASILY. STATES ARE SHOWN TO DIFFER IN THE COALITIONS DOMINATING CORRECTIONS ISSUES. IT WAS FOUND THAT BEING A STATE ELECTED OFFICIAL BRINGS ONE INTO THE NETWORK OF RECIPROCAL INFLUENCE AFFECTING DECISIONMAKING AT THE STATE LEVEL. ELITE MEMBERS WHO WERE ESPECIALLY WELL ACQUAINTED WITH POWERFUL GROUPS WHO HAD CONTACTED THEM AND EITHER INFLUENCED THEIR STANDS ON CORRECTIONS OR BEEN INFLUENCED BY THEM TENDED TO BE STATE ELECTED OFFICIALS AND THE BETTER EDUCATED AMONG THE RESPONDENTS. IT IS BELIEVED THAT STATE LEVEL DECISIONMAKING IN CORRECTIONS IS THUS DOMINATED BY ELECTED OFFICIALS AND LAWYERS. TABLES OF DATA AND FINDINGS ARE INCLUDED. (RCB)