THE DISCUSSION GROUPS WERE DESIGNED TO INVESTIGATE WHAT PERSONALITY CHANGES, IF ANY, TAKE PLACE AS A RESULT OF GROUP PARTICIPATION. THE PURPOSE OF THE PROJECT WAS TO DEVELOP A GROUP METHOD THAT COULD BE USED TO REDUCE THE RECIDIVISM RATES OF INMATES CURRENTLY SERVING TIME FOR VARIOUS OFFENSES. A NUMBER OF GROUPS WERE CONDUCTED FROM THE SUMMER OF 1965 THROUGH THE FALL OF 1967. THE BASIC APPROACH WAS TO CREATE A SET OF PERMISSIVE, ACCEPTING, NONRETALIATORY AND NONTHREATENING SMALL GROUPS FOR THE BENEFIT OF THE INMATES WHO COULD BRING PROBLEMS TO THE GROUPS FOR DISCUSSION AND COMMENT. GROUP PARTICIPANTS WERE COMPARED WITH NONGROUP PARTICIPANTS REGARDING PERSONALITY CHANGES. SPECIFIC HYPOTHESES WERE SET AND TESTED USING SEVERAL PERSONALITY TESTS--THE ROTTER SENTENCE COMPLETION TEST, INDEX OF ADJUSTMENT AND VALUES, AND THE MINNESOTA MULTIPHASIC PERSONALITY INVENTORY. RESULTS INDICATED THAT GROUP PARTICIPATION IS NOT SUPERIOR TO NONGROUP PARTICIPATION; INMATES PARTICIPATING IN A GROUP DISCUSSION PROGRAM DESIGNED TO INCREASE THEIR CIVILIAN ORIENTATION DO NOT SHOW SIGNIFICANT PERSONALITY CHANGES WHEN COMPARED TO NONGROUP PARTICIPANTS. A MASKING EFFECT MAY TAKE PLACE WHEN INMATES ARE ASSIGNED FOR GROUP PARTICIPATION. DATA ALSO SUGGESTED THAT INMATES IN THE MIDDLE RANGE OF MENTAL ABILITY AND EDUCATIONAL ACHIEVEMENT LEVEL CHANGED THE MOST ON SEVERAL VARIABLES, AND THAT QUALITATIVE AND QUANTATIVE CHANGES ARE A FUNCTION OF THE SPECIFIC INMATES IN THE GROUP PROGRAMS. TABULAR DATA AND A LIST OF 12 REFERENCES ARE INCLUDED. (MJW)