CitationHeiland, Frank W.; Korenman, Sanders; & Smith, Rachel A. (2018). On the sensitivity of estimates of the educational consequences of teenage childbearing: Sibling methods vs. instrumental variables. Annual Meeting of the Population Association of America. Denver, CO.
AbstractUnderstanding the contribution of childbearing to social disadvantages of teenage mothers requires estimates that both control for unobservables and generalize to teenage mothers. Using Add Health data, which oversamples siblings, and building on specifications from a widely cited study, we compare various sibling and Instrumental Variables (IV) estimates of the consequences of teenage childbearing on education. We find that the sibling and twin estimates are inconsistent with large adverse effects of teenage births. The IV-based estimates tend to be very sensitive to choice of instrument and specification. We discuss common concerns with sibling-based evidence (selectivity, power, generalizability, contamination) and argue that, despite their limitations, sibling estimates offer advantages and should be admitted as evidence on the consequences of teen childbearing.
Keyword(s)methodology teenage childbearing adolescence
Reference TypeConference proceeding
Book TitleAnnual Meeting of the Population Association of America
Series TitleSocioeconomic consequences of adolescent fertility and contraceptive access
Author(s)Heiland, Frank W.
Smith, Rachel A.