Instrument selection: The case of teenage childbearing and women's educational attainment
Recent research has identified situations in which instrumental variables (IV) estimators are severely biased and has suggested diagnostic tests to identify such situations. We suggest a number of alternative techniques for choosing a set of instruments that satisfy these tests from a universe of a priori plausible candidates, and we apply them to a study of the effects of adolescent childbearing on the educational attainment of young women. We find that substantive results are sensitive to instrument choice, and make two recommendations to the practical researcher: First, it is prudent to begin with a large set of potential instruments, when possible, and pare it down through formal testing rather than to rely on a minimal instrument set justified on a priori grounds. Second, the application of more restrictive tests of instrument validity and relevance can yield results very different from those based on less restrictive tests that produce a more inclusive set of instruments, and is the preferred, conservative approach when improper instrument choice can lead to biased estimates.
In order to set up a list of libraries that you have access to,
you must first login
or sign up.
Then set up a personal list of libraries from your profile page by
clicking on your user name at the top right of any screen.