The Buss–Perry Aggression Questionnaire (BP-AQ) is a 29-item, four-factor instrument that measures physical aggression, verbal aggression, anger, and hostility. A large number of competing versions of the original BP-AQ have been proffered in the literature, but testing with aggressive offenders is limited. The present study used confirmatory factor analysis to compare the fit of seven models with a sample of 246 men and women who pled guilty to misdemeanor aggressive offenses. The study found two similar parsimonious 12-item versions of the four-factor BP-AQ that outperformed the other models and fit the data well while retaining acceptable reliability. Taken together, superior psychometric properties, acceptable reliability, and reduced item-burden are properties that clearly support the usage of either of these 12-item versions. The relative merits of these shorter versions of the BP-AQ are examined and implications for their use in applied contexts and further research are considered.