The logistic regression model used in the study measured six factors in a violent encounter: the type of weapon, the demographic characteristics of victims and offenders, the circumstances of the encounter, the reported relationship between the offender and victim, the location of the encounter, and the time of the encounter. The study used 1995-2000 data from the FBI's National Incident-Based Reporting System (NIBRS), which has expanded the number of factors in violent crimes for which data are available. The number of sources for the NIBRS data were limited, however, with 9 States fully or partially participating in 1995 and 18 States fully or partially participating in 2000. The unit of analysis involved a single victim and a single offender. There were 262,493 incidents that involved 259,717 aggravated assaults and 2,776 homicides. The focus of the analysis was on factors that influenced whether the encounter resulted in an aggravated assault (nonlethal) or a homicide (lethal). Lethal outcomes were found to be more likely for older victims and male victims. Offenders who committed homicides were also more likely to be older and to be males. Acquaintance and stranger victim-offender relationships were less likely than family interactions to result in homicide. Firearms and knives were more likely than other weapons to result in death. Other factors measured also had some relationship to the outcome of the encounter; however, the type of weapon used and the circumstances of the encounter were the strongest predictors of a lethal outcome. 1 table, 10 notes, and 47 references