Should Instrumental Variables be Used as Matching Variables

By Jeffrey M. Wooldridge. 2016

143 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/ba7e757d93c15f1eecca73410c16c19594a76942

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Statistics and Probability

Citations in the corpus

Graphical criteria for efficient total effect estimation via adjustment in causal linear models (2019, background): “Conversely, if there is a path q from X to Y that is open given Z in every DAG D ∈ [G], then by the proof of Lemma 26 in Zhang (2008a) (see also the proof of Lemma 5.”
A Survey on Causal Inference (2020, methodology): “affected by the treatment assignment, should be excluded in the matching procedure [107]. Moreover, besides the post-treatment variables, researchers also suggest excluding the instrumental variables [93, 148], because they tend to amplify the bias of treatment effect estimator. 3.4 Tree-based Metho”
Directional Bias Amplification (2021, background): “Bias amplification is also studied in the causal statistics literature (Bhattacharya & Vogt, 2007; Wooldridge, 2016; Pearl, 2010; 2011; Middleton et al., 2016).”
A Systematic Study of Bias Amplification (2022, background): “Bias amplification has also been studied in the context of causal statistics [3, 30, 35, 36, 44], but that line of work has remained disparate from the study of bias amplification in machine learning.”
Cross-Border M&As and Innovative Activity of Acquiring and Target Firms (2016, background): “…as recent research suggests that matching on variables which satisfy IV assumptions increases the amount f inconsistency of matching estimators (Wooldridge, 2009). than patent counts as the outcome variable.39 In Table A8, results from dynamic Logit regressions, estimating the probability of a…”