Treatment Effects Using Inverse Probability Weighting and Contaminated Treatment Data: An Application to the Evaluation of a Government Female Sterilization Campaign in Peru

from event and focus on distinguishing whether a sterilization happened as result of the government campaign ( ) or without influence from the campaign ( . Making this distinction is still important because, as discussed earlier, the government campaign was allegedly directed at poor women, with little or no formal education (Coe, 2004), who had higher fertility rates and less knowledge of alternative contraceptive methods according to DHS II (1991-1992). We believe sterilizations likely had a different impact on this group of targeted women than on more educated or more affluent women who had knowledge about and access to a broader menu of contraceptive choices. To help fix ideas, let denote observed characteristics of a woman, such as education and place of residency (e.g. rural, urban). We parametrize the probability of being sterilized because of the campaign versus outside of the campaign as a function of linear indexes in , as shown in equations (2) and (3).