☆ 4.6 Article

When Should You Adjust Standard Errors for Clustering?*

QUARTERLY JOURNAL OF ECONOMICS

A CAUSAL BOOTSTRAP

Guido Imbens et al.

Summary: Bootstrap, initially introduced by The Jackknife, is a popular method for estimating variances and constructing confidence intervals by approximating the properties of estimators using the empirical distribution function of the sample. For causal estimands, uncertainty arises from the stochastic nature of the treatment received. A Bootstrap procedure for average treatment effect inference that considers this uncertainty is developed and compared to the classical Bootstrap, considering various designs.

ANNALS OF STATISTICS (2021)

添加到收藏夹

Article Economics

Bootstrap With Cluster-Dependence in Two or More Dimensions

Konrad Menzel

Summary: This proposal introduces a bootstrap procedure for data showing cluster-dependence in multiple dimensions, addressing the challenge of estimating limiting distributions under two-way clustering with adaptivity to different uniformity criteria. Important cases like regression inference, U- and V-statistics, subgraph counts for network data, and non-exhaustive samples of matched data are discussed in the paper.

ECONOMETRICA (2021)

添加到收藏夹

Article Economics