The study found that judges do take into account and weigh heavily legal variables when making decisions about pretrial detention. The severity of the current offense was a consistent predictor of both whether someone was detained and for how long. The number of prior arrests was also a significant and consistent predictor of these outcomes. Even after controlling for these legal factors in multivariate models, however, demographic characteristics such as age, gender, and race still influenced both pretrial detention and its length. Overall, the findings suggest that the use of an objective pretrial risk assessment could minimize the influence of extralegal variables in decision-making about pretrial detention. The study also found that the length of pretrial detention significantly increased the odds of conviction, but this does not necessarily mean that those who are detained longer will be convicted unfairly. Demographic characteristics were significantly related to many of the outcomes examined, including the likelihood of detention, the length of detention, time to adjudication, and the likelihood of conviction; for example, the odds of conviction were significantly higher for younger people and males in all models. This also supports the recommendation that a risk-assessment instrument be used as an objective method for assessing pretrial risk. 55 tables, 4 figures, and 15 references.