Cross-cultural Violence Risk Assessment: Adapting Risk Assessment for Institutional Violence Among Criminal Offenders in Mexico

Research has shown that violence risk assessment tools are less accurate with culturally and linguistically diverse populations. Despite the high rates of Hispanic/Latinx adults involved in the criminal justice system in the United States and Latin America, researchers have not examined whether the predictive accuracy of violence risk assessment measures would improve by the inclusion of culturally relevant risk factors. This prospective study was an investigation of the utility of a culturally adapted violence risk assessment measure and its predictive validity among a sample of male offenders in a Mexico City prison. Data collection was through clinical interviews and prison records review to examine the validity of a culturally adapted version of the Historical-Clinical-Risk Management-20V3 (HCR-20V3) in predicting institutional violence. Violent incidents for a 3-month period following each completed risk assessment were collected through records review, self-report exit interviews, and documentation provided by prison guards using the Modified Overt Aggression Scale (MOAS). It was hypothesized that the predictive validity of a culturally adapted version of the HCR-20V3 would be similar to that identified in previous literature with predominantly non-Hispanic white samples, that risk and protective factors that have been developed or adapted to reflect culturally relevant constructs would be significantly correlated with institutional violence, and that the culturally adapted HCR-20V3 would be a better predictor of violence than other predictors (e.g., psychopathy scores, age, violent index offense, or the original HCR-20V3 items). Area under the receiver operating characteristic (ROC) curve (AUC) and logistic regression analyses were used to examine the predictive validity of the culturally adapted HCR-20V3 and individual culturally relevant risk and protective factors as predictors of institutional violence. The study findings showed that the culturally adapted HCR-20V3 was able to predict institutional violence with a high degree of accuracy and that the culturally adapted HCR-20V3 Future Violence/Case Prioritization score provided incremental validity for predicting violence when compared with other empirically supported predictors. Additionally, several culturally relevant risk factors, such as the Indirect and Direct Violence Exposure, Familismo (Negative Influence), and Machismo Attitudes items, significantly predicted institutional violence. Although not all of the study hypotheses were supported, the results suggest that culture plays an important role in understanding and estimating violence risk