The study found that FTA was slightly more likely among cases with a defendant released on recognizance (ROR) than among cases with a defendant released on bail. The FTA rate was 17 percent for ROR compared to 14 percent for bail. The multivariate logistic regression model confirmed that for each bail amount range over $1.00, a reduction in the predicted probability of FTA was statistically significant, compared to the probability of FTA among ROR defendants. Very low bail had nearly the same probability of FTA as higher bail. Even bail between $50 and $500 lowered the likelihood of FTA compared to the 17-percent predicted probability of FTA for ROR cases. The defendants who were release through bondsmen were the only ones in the research sample who were monitored or supervised in any way, and this supervision did not produce lower FTA rates than cash bail. The one area in which commercial bonds were found to have a small positive effect was in returning defendant to court within 30 days, once they had missed a scheduled date. The study focused on the impact of commercial bonds because the poor are the hardest hit by bondsmen's fees. Defendants with more money are able to post cash bail instead, which is refunded at the end of the case, less a 3 percent fee if the defendant was convicted. In New York City, commercial bonds do nothing to lower FTA rates and are responsible for the release of a disproportionate number of violent, high-bail defendants. 21 tables, 8 figures, 41 references, and appended description of the multivariate statistical procedure used in the study