Federal Probation Volume: 75 Issue: 2 Dated: September 2011 Pages: 78-87
This article reports on a meta-analysis intended to identify from previous research what risk factors are statistically associated with various measures of pretrial failure (re-arrest, failure to appear in court, new crime, and any failure to comply with pretrial-release conditions).
The findings of previous studies on factors linked to failure on pretrial release yielded very few correlations that were strong predictors of pretrial failures. Those risk factors with the strongest correlations that were also in the expected direction are primarily static indicators, such as prior convictions, prior felonies, prior misdemeanors, prior failure to appear, and juvenile arrests. Scores on pretrial risk assessments were found to be significant, suggesting that higher scores on pretrial risk assessments correlated significantly with pretrial failure outcomes. Although there were insignificant predictors of pretrial failures, several factors for which data were obtained to determine pretrial release success were either found to be in the wrong direction or were consistently insignificant for predicting pretrial-release failures. These are residency, injury to a victim, using a weapon, and abusing alcohol. Although the authors advise that this study should be considered preliminary in the identification of pretrial risk predictors, several policy implications are offered for consideration. First, the administration of pretrial risk tools should continue. Second, residency, weapons, alcohol use, and victim injury might still be risk items for pretrial failure; however, future research should explore whether or not these factors are consistently related to nonappearance at court or risk to public safety. Third, studies of pretrial risk predictors or the validation of pretrial risk assessment tools should be published or publicly disseminated. Fourth, research on pretrial risk factors and assessment tools should provide the necessary statistics, so as to allow for the calculation of effect sizes. 12 tables, 20 references, and appended supplementary information on the methodology of the meta-analysis
United States of America