"Donohue and Levitt use no data on abortion prior to 1973. Their analysis of arrests by single year of age, for instance, pertains to birth cohorts born between 1961 and 1981 where approximately 60 percent of the state/ age / cohort cells are assigned an abortion ratio of zero. However, demographers have concluded that most legal abortions in the early 1970s replaced illegal abortions (Tietze 1973; Sklar and Berkov 1974). If the underreporting of abortion were random among states, their estimates would be biased downward. As I show below, however, the measurement error is negatively correlated with the true abortion rate in 1972 and thus the direction of the bias is unknown."