A unique dataset is analyzed to investigate the effect of a criminal suspect’s prior criminal record on the probability of arrest. Multivariate logistic regression results show that a criminal suspect with a prior criminal record is approximately 29 times more likely than a suspect without a criminal record to be arrested by police. While findings also reveal that Black suspects and Black suspects with a prior criminal record do not have an enhanced proclivity of arrest, Black suspects with a prior criminal record who target White victims are almost three times more apt to be arrested. When juxtaposed with the finding in the baseline model of a substantive relationship between a suspect’s race and the likelihood of arrest absent the control for prior criminal record, our results suggest that any correlation evinced between a criminal suspect’s race and the likelihood of arrest without controlling for the suspect’s prior criminal history may be spurious due to omitted variable bias.

分析了唯一的数据集，以调查犯罪嫌疑人先前的犯罪记录对逮捕概率的影响。多元logistic回归结果显示，具有先前犯罪记录的犯罪嫌疑人比没有犯罪记录的犯罪嫌疑人被警察逮捕的可能性大约高29倍。虽然调查结果还显示，黑人犯罪嫌疑人和有犯罪记录的黑人嫌疑人的逮捕倾向没有增强，但针对白人受害者的有犯罪记录的黑人犯罪嫌疑人的犯罪几率要高出将近三倍。如果将基线模型中的发现与犯罪嫌疑人的种族和在没有事先犯罪记录控制的情况下逮捕的可能性之间存在实质性关系并列，