Bounding treatment effects: A command for the partial identification of the average treatment effect with endogenous and misreported treatment assignment

• Autores: Ian McCarthy, Daniel L. Millimet, Manan Roy
• Localización: The Stata journal, ISSN 1536-867X, Vol. 15, Nº. 2, 2015, págs. 411-436
• Idioma: inglés
• DOI: 10.1177/1536867x1501500205
• Texto completo no disponible (Saber más ...)
• Resumen

  • We present a new command, tebounds, that implements a variety of techniques to bound the average treatment effect of a binary treatment on a binary outcome in light of endogenous and misreported treatment assignment. To tighten the worst case bounds, the monotone treatment selection, monotone treatment response, and monotone instrumental-variable assumptions of Manski and Pepper (2000, Econometrica 68: 997–1010), Kreider and Pepper (2007, Journal of the American Statistical Association 102: 432–441), Kreider et al. (2012, Journal of the American Statistical Association 107: 958–975), and Gundersen, Kreider, and Pepper (2012, Journal of Econometrics 166: 79–91) may be imposed. Imbens–Manski confidence intervals are provided.