Work

Further Evidence on the Internal Validity of the Early Legal Access Research Design

Year: 2013

Type: article

Abstract: Randomized controlled experiments are the gold standard of causal analysis, but they are unavailable to answer many questions of policy interest.Access to the Pill or abortion, for instance, was never explicitly randomly assigned, so quasi-experiments provide an important path for understanding their effects.We agree with some of Joyce's arguments regarding abortion.But his arguments regarding the Pill lose site of the forest for the trees and distract from the major policy questions at stake.His comments on the Pill focus narrowly on one genre of studies-those using as a research design changes in the age of consent for contraception ("early legal access to the Pill," ELA)-and ignore the larger body of quasi-experimental evidence on the broad benefits of reproductive health policy.In this response, we address his four critiques of the ELA literature, describe the assumptions necessary for them to matter in practice, and present historical and empirical evidence on their relevance.Joyce's first critique is that dating legal changes in Pill access is difficult.We agree and have made this point ourselves (Bailey, Hershbein, & Miller, 2012; Bailey, Guldi, Davido, & Buzuvis, 2011).The difficulty in coding these laws today does not imply they did not matter.Women in the 1960s would have had no trouble figuring out whether they could obtain the Pill (by being denied a prescription, for instance).Moreover, misclassification of legal changes cannot explain the literature's results.With a binary independent variable (ELA), perfect misclassification leads to finding the opposite of the expected effects (Jakubson, 1986).In the presence of more modest misclassification, the literature would tend to understate the Pill's effects (the reverse of Joyce's argument).In practice, this problem appears minor.Figure 1 shows that the effects of early access to the Pill are statistically indistinguishable between Bailey's legal coding (2006), the earliest published legal coding, and the most recent revision in Myers (2012).Access to the Pill between 18 and 21 affects the likelihood of giving birth at precisely those ages, which lends credibility to the research design.

Source: Journal of Policy Analysis and Management

DOI: 10.1002/pam.21716

Authors: Martha J. Bailey, Melanie Guldi, Brad J. Hershbein

Institutions: University of Central Florida, National Bureau of Economic Research, University of Michigan, W.E. Upjohn Institute for Employment Research

Language: English

Topic: Reproductive Health and Contraception

Subfield: Public Health, Environmental and Occupational Health

Field: Medicine

Domain: Health Sciences

SDGs: Good health and well-being, Peace, justice, and strong institutions

Open Access status: bronze

Funders: University of Michigan, College of Pharmacy, University of Michigan, Eunice Kennedy Shriver National Institute of Child Health and Human Development