Similar to previous studies for other cities, the findings showed significant differences between gang and nongang homicides. Just as in various other cities, gang homicides in Newark were more likely than nongang homicides to occur in public settings, involve firearms and multiple suspects, and have minority youth male victims and/or suspects. The authors argue that the findings support Decker's (1996) collective behavior theory of gang homicides. This theory explains gang violence, including homicide, to be an expressive result of collective behavior. Decker describes a seven-step process whereby "threats" from symbolic enemies trigger an escalating response that peaks in violence and retaliation. The authors of the current study recommend that the research models used in Newark be used in other cities. The homicide investigation files of the Newark Police Department showed that 417 homicides occurred in Newark over the study period. In distinguishing gang homicides, the Newark Police Department used the "Los Angeles definition" of gang homicides, i.e., incidents in which one of the actors was a gang member at the time of the homicide. After excluding 75 homicide cases because it was not known whether there was any gang involvement, 342 cases remained for analysis. Of these, 137 (40 percent) were coded as gang-related. The dependent variable was the homicide incident, and the independent variables pertained to incident characteristics and the demographic characteristics of the victims and suspects by homicide type. 3 tables, 7 notes, and 61 references