THE SURVEY OBJECTIVE WAS TO LEARN WHAT STATES ARE DOING IN THE AREA OF INMATE POPULATION PROJECTIONS RATHER THAN TO JUDGE OR RANK THESE EFFORTS. OF 44 RESPONDING STATES, 20 FURNISHED DOCUMENTS AND 32 INDICATED THEY PUBLISH PROJECTIONS; 21 STATES HAD USED LINEAR REGRESSION, 7 STATES HAD USED MULTIPLE REGRESSION, AND 6 STATES HAD USED SIMULATION MODELS. SINCE IT WAS WIDELY RECOGNIZED THAT PAST GROWTH DOES NOT DETERMINE FUTURE GROWTH, VARIOUS LINEAR REGRESSION METHODS BASED ON PAST INMATE POPULATION FIGURES WERE NOT SATISFACTORY. MULTIPLE REGRESSION WAS EMPLOYED IN SEVERAL STATES TO DETERMINE WHAT COMBINATION OF FACTORS HAS THE GREATEST PREDICTIVE ABILITY. IN COLORADO, FOR EXAMPLE, THE FOLLOWING FACTORS ARE USED TO PREDICT NEW COURT COMMITMENTS: UNEMPLOYMENT RATE FOR THE PRECEDING QUARTER; AVERAGE PERCENTAGE OF YEARLY COMMITMENTS RECEIVED EACH QUARTER; AND POPULATION AT RISK. WISCONSIN EMPLOYS MULTIPLE REGRESSION AND USES AGE-SPECIFIC CENSUS DATA, ECONOMIC FACTORS LIKE EMPLOYMENT AND INCOME, AND ARREST RATES AND CONVICTION DATA. THE DEGREES OF SUCCESS REPORTED BY STATES USING MULTIPLE REGRESSION IS NOT ALL THAT ENCOURAGING, HOWEVER, WITH RESPECT TO RELIABILITY. SOME STATES APPROACH POPULATION PROJECTIONS WITH THE AID OF SIMULATION MODELS. THE PURPOSE OF THESE MODELS IS TO REPRODUCE THE CRIMINAL JUSTICE SYSTEM'S OPERATION OR PART OF THAT SYSTEM. THE SIMULATION MODEL IN GEORGIA ATTEMPTS TO SIMULATE OFFENDER MOVEMENT THROUGH THE PRISON SYSTEM. VIRGINIA DIVIDES THE INMATE POPULATION INTO FOUR GROUPS (MISDEMEANANTS, FELONS, CURRENTLY CONFINED IN STATE INSTITUTIONS, CONVICTED FELONS IN LOCAL JAILS WHO ARE AWAITING ADMISSION TO THE STATE SYSTEM, AND FELONS EXPECTED TO BE COMMITTED DURING THE PROJECTION TIME FRAME). NORTH CAROLINA HAS DEVISED A COMPLICATED SERIES OF MODELS THAT REFLECT CORRECTIONAL SYSTEM FLOW (SIMPLE FLOW, MARKOV DETERMINISTIC, SYSTEMS-DYNAMICS, AND DISCRETE EVENT DIGITAL SYSTEM SIMULATION MODELS). A NOTEWORTHY FEATURE OF MARYLAND'S SIMULATION TECHNIQUE IS THAT THE ANALYSIS OF ARREST RATES AND THE PROBABILITY OF INCARCERATION IS BROKEN DOWN INTO SPECIFIC CRIME TYPES. CORRECTIONAL PLANNERS IN INDIANA INCLUDE SUCH FACTORS AS SOCIETAL ATTITUDES AND CONDITIONS, POPULATION AND ECONOMIC CONDITIONS, CRIME AND ARREST RATES, IMPLICATIONS OF THE STATE CRIMINAL CODE, USE OF PROBATION AND OTHER COMMUNITY ALTERNATIVES, AND PAROLE POLICIES IN THEIR SIMULATIONS. MOST STATES REPORT THERE ARE CONSTRAINTS IN MAKING POPULATION PROJECTIONS, E.G., LACK OF HIGH-QUALITY DATA, ABSENCE OF SPECIALLY-TRAINED STAFF, AND POLICY CHANGES. DATA ON INMATE POPULATION PROJECTION TECHNIQUES AND FACTORS USED BY STATES IN MULTIPLE REGRESSION METHODS ARE TABULATED.