THE METHOD OF CONSTRUCTION IS CALLED SYSTEMS DYNAMIC SIMULATION MODELING, WHICH PERMITS SYMBOLIC REPRESENTATION OF SYSTEM STRUCTURES AND OF DYNAMIC CHANGES IN PROCESS AND OUTCOMES. THE CRIMINAL JUSTICE SYSTEM MODEL CONSISTS OF SUBSYSTEMS FOR THE AT-RISK (OF COMMITTING A CRIME) AGE GROUP, OFFENDERS AND OFFENSES, AND COURTS AND CORRECTIONS; POPULATION IS TREATED AS AN EXOGENOUS VARIABLE. DATA FROM A VARIETY OF SOURCES ARE USED TO CONSTRUCT THE MODEL SUBSYSTEMS. SIMULATION RUNS BASED ON FIVE VARIATIONS OF THE MODEL SUGGEST THE FOLLOWING CONCLUSIONS ABOUT THE CRIMINAL JUSTICE SYSTEM: (1) THE LONG-TERM RESULT OF IMPROVING LAW ENFORCEMENT EFFICIENCY AND THEREBY INCREASING THE PROPORTION OF OFFENDERS APPREHENDED WILL BE AN INCREASE IN THE CRIME RATE, (2) INCREASES IN PRISON SENTENCES WILL PLACE A LARGE FISCAL BURDEN ON TAXPAYERS WITHOUT REDUCING THE CRIME RATE, AND (3) THE MOST EFFECTIVE POINT OF CONTROL IN THE CRIMINAL JUSTICE SYSTEM IS THE FIRST OFFENSE--THE JUNCTURE AT WHICH NONOFFENDERS BECOME OFFENDERS. THESE CONCLUSIONS ARE NOT INTENDED TO BE PREDICTIVE. NOR IS THE MODEL TO BE USED FOR POLICY PURPOSES. IT IS INTENDED TO FURTHER THE DEVELOPMENT OF MODELS AND MODELING METHODS, AND TO PROVIDE A MEANS OF ASKING 'WHAT IF' QUESTIONS ABOUT THE CRIMINAL JUSTICE SYSTEM. DETAILS OF THE MODEL, FLOWCHARTS, A LIST OF REFERENCES, AND DISCUSSIONS OF MODELS, SYSTEMS, AND SIMULATION ARE INCLUDED. (LKM)