The current study evaluated a range of social influences including misdemeanor arrests, drug arrests, cocaine consumption, alcohol consumption, firearm availability, and incarceration that may be associated with changes in gun-related homicides by racial/ethnic group in New York City (NYC) from 1990 to 1999. Using police precincts as the unit of analysis, we used cross-sectional, time series data to examine changes in Black, White, and Hispanic homicides, separately. Bayesian hierarchical models with a spatial error term indicated that an increase in cocaine consumption was associated with an increase in Black homicides. An increase in firearm availability was associated with an increase in Hispanic homicides. Last, there were no significant predictors for White homicides. Support was found for the crack cocaine hypotheses but not for the broken windows hypothesis. Examining racially/ethnically disaggregated data can shed light on group-sensitive mechanisms that may explain changes in homicide over time. (Published Abstract)