THE FIRST MISCONCEPTION IS THAT SOME PROGRAMS ARE NOT SUITABLE FOR EVALUATION. HERE, DIFFICULTY HAS BEEN MISTAKEN FOR IMPOSSIBILITY WHEN ELABORATE EVALUATION TECHNIQUES AND PROPER RESEARCH DESIGN COULD HAVE SOLVED RESEARCH PROBLEMS. A SECOND ERRONEOUS BELIEF IS THAT AN INFORMAL PROGRAM EVALUATION CAN RESULT IN SOUND CONCLUSIONS, WHILE A THIRD IS THAT RECIDIVISM PERCENTAGES BY THEMSELVES ARE SUFFICIENTLY INFORMATIVE OF PROGRAM EFFECTIVENESS TO AND IN MAKING JUDGMENTS APART FROM COMPARATIVE DATA. A FOURTH MISCONCEPTION CONCERNS RANDOM ASSIGNMENT. MANY PRACTITIONERS FEAR RANDOM ASSIGNMENT BECAUSE THEY PRESUME IT TO MEAN THAT ANY INMATE CAN BE ASSIGNED TO AN EXPERIMENTAL GROUP REGARDLESS OF INDIVIDUAL CHARACTERISTICS WHICH MAY OR MAY NOT SUIT THAT INMATE FOR A PROGRAM. IN REALITY, RESEARCHERS USE CRITERIA FOR SELECTING A GENERAL STUDY POPULATION BEFORE OBTAINING A RANDOM SAMPLE FROM THAT GENERAL POPULATION. ANOTHER FAULTY BELIEF IS THAT PROGRAM EVALUATIONS PRODUCE EITHER/OR RESULTS--EITHER THE PROGRAM IS A SUCCESS OR IT IS A FAILURE. BELIEFS THAT PROGRAM EVALUATIONS ARE COSTLY AND THAT RESEARCH, WHEN CONDUCTED, SHOULD MAKE MAJOR BREAKTHROUGHS TO BE WORTHWHILE ARE ALSO ERRONEOUS. BESIDES BEING VULNERABLE TO THESE MISCONCEPTIONS, CORRECTIONAL RESEARCH SUFFERS FROM MISUSE OF STATISTICS. THE FOLLOWING ARE COMMON RESEARCH PROBLEMS: POORLY IDENTIFIED STATISTICS CONCERNING THE 'AVERAGE' INMATE, WHOSE CHARACTERISTICS MAY BE VARIOUSLY MEASURED BY THE ARITHMETIC MEAN, THE MODE, OR THE MEDIAN; A CONFUSION OF THIS 'AVERAGE' TYPE WITH THAT WHICH IS NORMAL OR DESIRABLE; AND CONFUSION BETWEEN THE CORRELATIONAL AND CAUSATIVE RELATIONSHIPS OF PHENOMENA. TO AVOID MISCONCEPTIONS AND MISTAKES, RESEARCHERS MUST TAKE GREAT CARE IN EXAMINING DATA AND THEIR RELATIONSHIPS, AND PRACTITIONERS SHOULD CRITICALLY STUDY THE RESEARCHER'S IDENTITY AND LOOK INTO POSSIBLE VESTED INTERESTS IN ADDITION TO BEING COGNIZANT OF THE FOREGOING DIFFICULTIES. FOOTNOTES ARE INCLUDED. (MRK)