This article disaggregates Donohue and Levitt's (DL's) national panel-data models to the state level and shows that high concentrations of teenage abortions in a handful of states drive all of DL's results in their 2001, 2004, and 2008 articles on crime and abortion. These findings agree with previous research showing teenage motherhood is a major maternal crime factor, whereas unwanted pregnancy is an insignificant factor. Teenage abortions accounted for more than 30% of U.S. abortions in the 1970s, but only 16% to 18% since 2001, which suggests DL's panel-data models of crime/arrests and abortion were outdated when published. The results point to a broad range of future research involving teenage behavior. A specific means is proposed to reconcile DL with previous articles finding no relationship between crime and abortion.