Simple Tests for Selection Bias: Learning More from Instrumental Variables

Black, Dan A.; Joo, Joonhwi; Lalonde, Robert; Smith, Jeffrey A.; Taylor, Evan J.

doi:10.2139/ssrn.2663776

Cited by 9 publications

(6 citation statements)

References 31 publications

(43 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…That is, the same testable implication as considered in Theorem 1 also arises when imposing conditional IV validity and testing the selection-on-observables assumption w.r.t. treatment, see also Black et al (2015) for a related study. Taking in some sense the opposite approach, one may alternatively assume a treatment that is unconfounded to test IV validity, which again entails the same testable implication.…”

Section: Testable Implication For Identificationmentioning

confidence: 99%

“…This paper connects to several strands of the causal inference literature. Most closely related are studies assuming an instrument that is conditionally valid given covariates, in order to test the selection-on-observables assumption on the treatment based on the very same condition as in this paper, see de Luna and Johansson (2014) and Black et al (2015). Angrist (2004), Brinch et al (2017), and Huber (2013 consider related tests when instrument validity is assumed to hold unconditionally, i.e.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Testing the identification of causal effects in data

Huber¹

2022

Preprint

View full text Add to dashboard Cite

This study demonstrates the existence of a testable condition for the identification of the causal effect of a treatment on an outcome in observational data, which relies on two sets of variables, namely observed covariates to be controlled for and a suspected instrument.Under a causal structure commonly found in empirical applications, the testable conditional independence of the suspected instrument and the outcome given the treatment and the covariates has two implications. First, the instrument is valid, i.e. it does not directly affect the outcome (other than through the treatment) and is unconfounded conditional on the covariates. Second, the treatment is unconfounded conditional on the covariates such that the treatment effect is identified. We suggest tests of this conditional independence based on doubly robust estimators and investigate their finite sample performance in a simulation study. We also apply our testing approach to the evaluation of the impact of fertility on female labor supply when using the sibling sex ratio of the first two children as supposed instrument, which by and large points to a violation of our testable implication, at least for the moderate set of socio-economic covariates considered.

show abstract

Section: Testable Implication For Identificationmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Testing the identification of causal effects in data

Huber¹

2022

Preprint

View full text Add to dashboard Cite

show abstract

“…as a test of no selection bias in the literature on instrumental variables (Black et al 2015;Brinch, Mogstad, and Wiswall 2017). While T * is exogenous under Assumption 2, we relax this assumption below to admit some forms of endogeneity.…”

Section: The Model and Identificationmentioning

confidence: 99%

Correcting for Misclassified Binary Regressors Using Instrumental Variables

Haider¹,

Stephens²

2020

View full text Add to dashboard Cite

the 2019 Society of Labor Economists meetings, and the Michigan-Michigan State-Western University Labo(u)r Day Conference for helpful comments and suggestions. The views expressed herein are those of the authors and do not necessarily reflect the views of the National Bureau of Economic Research.NBER working papers are circulated for discussion and comment purposes. They have not been peer-reviewed or been subject to the review by the NBER Board of Directors that accompanies official NBER publications.

show abstract

“…As discussed in Angrist (2004), the LATE is directly externally valid, i.e., corresponds to the ATE when either all mean potential outcomes are homogeneous across compliance types, or at least the average effects. For testing the equality of mean potential outcomes across treated compliers and always takers as well as across nontreated compliers and never takers, see Angrist (2004), de Luna and Johansson (2014), Huber (2013), andBlack et al (2015). See also Donald et al (2014b) for a related, but yet different testing approach.…”

Section: Instrumental Variablesmentioning

confidence: 99%

An introduction to flexible methods for policy evaluation

Huber¹

2019

Preprint

View full text Add to dashboard Cite

This chapter covers different approaches to policy evaluation for assessing the causal effect of a treatment or intervention on an outcome of interest. As an introduction to causal inference, the discussion starts with the experimental evaluation of a randomized treatment. It then reviews evaluation methods based on selection on observables (assuming a quasi-random treatment given observed covariates), instrumental variables (inducing a quasi-random shift in the treatment), difference-in-differences and changes-in-changes (exploiting changes in outcomes over time), as well as regression discontinuities and kinks (using changes in the treatment assignment at some threshold of a running variable). The chapter discusses methods particularly suited for data with many observations for a flexible (i.e. semior nonparametric) modeling of treatment effects, and/or many (i.e. high dimensional) observed covariates by applying machine learning to select and control for covariates in a data-driven way. This is not only useful for tackling confounding by controlling for instance for factors jointly affecting the treatment and the outcome, but also for learning effect heterogeneities across subgroups defined upon observable covariates and optimally targeting those groups for which the treatment is most effective.

show abstract

Simple Tests for Selection Bias: Learning More from Instrumental Variables

Cited by 9 publications

References 31 publications

Testing the identification of causal effects in data

Testing the identification of causal effects in data

Correcting for Misclassified Binary Regressors Using Instrumental Variables

An introduction to flexible methods for policy evaluation

Contact Info

Product

Resources

About