Randomization Tests Under an Approximate Symmetry Assumption

By Ivan A. Canay, Joseph P. Romano, A. Shaikh. 2017

123 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Statistics and Probability

View Paper: https://www.semanticscholar.org/paper/169fdceb057b381b4efe06d71df71f5c0db11dd0

How researchers see this paper

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

In Artificial Intelligence

In Statistics and Probability

Citations in the corpus

Inference in Regression Discontinuity Designs with a Discrete Running Variable (2016, methodology): “5Using alternative methods for constructing CIs suggested recently in the literature on inference with a small number of clusters (e.g. Cameron et al., 2008; Canay et al., 2015; Imbens and Kolesár, 2016) does not lead to reliable inference in this case: as we argued above, even if standard errors we”
Placebo Tests for Synthetic Controls (2017, methodology): “Following Canay et al. (2014), such test would be asymptotically valid if the test statistics are based on a function of the data that exhibits approximate symmetry.”
The Standard Errors of Persistence (2019, unknown): “9 Such tuning “parameters” can be choice of large clusters, terms in an approximation, or a reference correlation function, e.g. Conley 1999, Ibragimov and Müller (2010), Kim and Sun (2011), Canay, Romano and Shaikh (2017), and Müller and Watson (2022, 2023).”
Approximate permutation tests and induced order statistics in the regression discontinuity design (2015, methodology): “This is possible in our context due to the recent asymptotic framework developed by Canay et al. (2017) for randomization tests, although we introduce novel modifications to make it work in the RDD setting — see Section 4.2.”
Measuring Success in Education: The Role of Effort on the Test Itself (2017, methodology): “and Shanghai, which we calculate using a randomization test (Canay et al., 2017).”
Synthetic Control and Inference (2017, methodology): “Although we will later use the estimator in (7) in our Monte Carlo simulations, we do not need to restrict ourselves to this particular estimator.”
Asymptotic theory for clustered samples (2019, unknown): “A sufficient condition for the CLT (9) is that for all ε > 0”
Cluster-robust inference: A guide to empirical practice (2022, background): “The results of Canay et al. (2021) and MacKinnon and Webb (2020b) strongly suggest that homogeneity across clusters is more important for RI than for the”
The Wild Bootstrap with a “Small” Number of “Large” Clusters (2019, background): “This distinction prevents the application of results in Canay et al. (2017). In fact, sn and ŝn do not even tend in distribution to the same limit.”
Covariate-adjusted Fisher randomization tests for the average treatment effect (2020, background): “There is an increasing interest in economics and related fields to use frts to analyze various types of empirical data (Freedman and Lane 1983; Kennedy 1995; Cattaneo et al. 2015; Canay et al. 2017; Ganong and Jäger 2018; Athey et al. 2018; Bugni et al. 2018; Young 2019; Heckman and Karapakula 2019;”
Placebo inference on treatment effects when the number of clusters is small (2018, methodology): “h. I consider (i) placebo inference as in Algorithm2.5, (ii) two-sample inference with theIbragimov and Muller(2016) ttest, (iii) randomized and non-randomized inference with the permutation ttest ofCanay et al. (2014), (iv) cluster-robust wild bootstrap inference as described inCameron et al.(2008)”
Testing continuity of a density via g-order statistics in the regression discontinuity design (2018, background): “The test φ(Sn) could be alternatively characterized as an “approximate” randomization test, see Canay et al. (2017) for a general description of such tests.”
What’s trending in difference-in-differences? A synthesis of the recent econometrics literature (2022, unknown): “When in fact the number of time periods is small, as frequently occurs in DiD applications, one can still use some of these methods, but the underlying assumptions are stronger — see, e.g., Remark 4.5 and Section 4.2 of Canay et al. (2017).”