Data were obtained from records that covered the period 1980 through 1992. For each of these years, a random sample of 3,637 persons who had their first police contact during that year was drawn from the computerized central file at Israeli Police Headquarters. A total of 40,007 individuals with a total of 97,000 records constituted the study population. The findings identified which criminal cases were most likely to be terminated prior to adjudication and which cases, once adjudicated, were most likely to result in convictions. A major focus of the analysis was the nationality of the defendant (Arab or Jew), while the effect of other variables, such as the type of offense and the time period, were controlled. The criminal justice system was found to be less discriminating at the early stages of the criminal process, but as the offender moved on in the process, the chances that nationality would play a part in case decision-making increased. In all offense categories and in all profiles examined, Arabs had a greater chance of conviction than Jews. The pattern of discrimination toward Arabs did not apparently increase with the overall increased probability for conviction, nor did the rate of case termination change over the years by nationality. This suggests that the greater punitiveness exhibited by the system and the proactive policing practices were not driven by considerations of nationality or ethnicity but by genuine (although not empirically justifiable by crime statistics) concern with the crime problem in Israel. 6 tables, appended coding format, and 29 references