Article
Economics
James G. MacKinnon, Morten Orregaard Nielsen, Matthew D. Webb
Summary: The study examines two cluster-robust variance estimators for regression models with clustering in two dimensions and provides conditions for asymptotically valid inferences based on each. It also introduces several wild bootstrap procedures with conditions for their asymptotic validity for each type of t-statistic. Simulation results suggest that using certain bootstrap procedures with one of the t-statistics generally performs well, and an empirical example shows substantial differences between bootstrap inferences and conventional ones.
JOURNAL OF BUSINESS & ECONOMIC STATISTICS
(2021)
Article
Economics
Carlos Lamarche, Thomas Parker
Summary: This work proposes a new method for addressing statistical inference in penalized quantile regression for longitudinal data, which provides an asymptotically valid approximation of the estimator's distribution and imposes no restrictions on individual effects. Simulation studies demonstrate that the method has accurate small sample behavior and is easy to implement. The authors also illustrate the new approach using U.S. Census data.
JOURNAL OF ECONOMETRICS
(2023)
Article
Economics
James G. MacKinnon, Morten Orregaard Nielsen, Matthew D. Webb
Summary: This paper provides a comprehensive guide to empirical practice by using recent econometric theory and simulation evidence to bridge the gap between theory and practice in cluster-robust inference methods. It includes an empirical analysis of the minimum wage's effects on teenage labor supply using individual data.
JOURNAL OF ECONOMETRICS
(2023)
Article
Economics
James G. MacKinnon, Morten Orregaard Nielsen, Matthew D. Webb
Summary: The majority of empirical research assumes known clustering structure, while this study proposes tests for determining the appropriate level of clustering in regression models. Both asymptotic and wild bootstrap implementations are provided, and simulations suggest excellent performance of the bootstrap tests. These tests lead to sensible inferences as shown in an empirical example.
JOURNAL OF ECONOMETRICS
(2023)
Article
Computer Science, Interdisciplinary Applications
Adam Loy, Jenna Korobova
Summary: This paper introduces the lmeresampler package for bootstrapping nested linear mixed-effects models, providing bias correction, adjusted standard errors, and confidence intervals. It also demonstrates the use of bootstrap resampling for diagnosing this model class. In addition, lmeresampler makes it easy to construct interval estimates of functions of model parameters.
Article
Statistics & Probability
Shonosuke Sugasawa
Summary: The article introduces a flexible and interpretable modeling approach, called grouped heterogeneous mixture modeling, for clustered data. The model uses mixtures of latent conditional distributions to model cluster-wise conditional distributions, assuming that clusters are divided into a finite number of groups with the same mixing proportions within the same group. Structured grouping strategies and algorithms for computing the maximum likelihood estimator are proposed.
JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION
(2021)
Article
Economics
James G. MacKinnon, Morten Orregaard Nielsen, Matthew D. Webb
Summary: We provide computationally attractive methods to obtain jackknife-based cluster-robust variance matrix estimators (CRVEs) for linear regression models estimated by least squares. We also propose several new variants of the wild cluster bootstrap, which involve these CRVEs, jackknife-based bootstrap data-generating processes, or both. Extensive simulation experiments suggest that the new methods can provide much more reliable inferences than existing ones in cases where the latter are not trustworthy, such as when the number of clusters is small and/or cluster sizes vary substantially. Three empirical examples illustrate the new methods.
JOURNAL OF APPLIED ECONOMETRICS
(2023)
Article
Computer Science, Interdisciplinary Applications
Renwen Luo, Jianxin Pan
Summary: Joint modelling of mean-covariance structure is important in clustered data analysis. Existing methods have limitations in assuming natural order in responses and modeling transformed parameters. The proposed data-driven method is flexible, interpretable, and works on original correlation coefficients and variances without the need for natural order.
COMPUTATIONAL STATISTICS & DATA ANALYSIS
(2022)
Article
Mathematical & Computational Biology
Megha Joshi, James E. Pustejovsky, S. Natasha Beretvas
Summary: Meta-analytic reviews in social science research often include multiple effect size estimates per study, leading to dependence in the estimates. An alternative method called cluster wild bootstrapping has been evaluated and shown to maintain adequate error rates and provide more power compared to existing small-sample correction methods, particularly for multiple-contrast hypothesis tests.
RESEARCH SYNTHESIS METHODS
(2022)
Article
Engineering, Environmental
Carlo Correa Solci, Valderio Anselmo Reisen, Paulo Canas Rodrigues
Summary: This paper proposes a generalization of the local bootstrap for periodogram statistics when weakly stationary time series are contaminated by additive outliers. The robust bootstrap periodogram is implemented in the Whittle estimator to obtain confidence intervals for the parameters of a time series model. The results have shown that the robust estimator is resistant to additive outlier contamination and produces confidence intervals with coverage percentages closer to 95% and lower amplitudes than the ones obtained with the classical estimator, even for small percentages and magnitudes of outliers.
STOCHASTIC ENVIRONMENTAL RESEARCH AND RISK ASSESSMENT
(2023)
Article
Psychology, Multidisciplinary
Daniel McNeish
Summary: This article introduces different methods for dealing with clustering issues in psychological data and emphasizes that the choice of statistical method should be guided by the research question. The article provides a user-friendly resource to help psychologists understand and apply these methods.
PSYCHOLOGICAL METHODS
(2023)
Article
Statistics & Probability
Ting Ye, Jun Shao, Hyunseung Kang
Summary: The study provides a theoretical characterization of popular estimators in Mendelian randomization under many weak instruments. It introduces a debiased IVW estimator that is robust to many weak instruments and does not require screening. Two instrument selection methods are also presented to improve the efficiency of the new estimator when a selection dataset is available.
ANNALS OF STATISTICS
(2021)
Article
Economics
J. Krampe, E. Paparoditis, C. Trenkler
Summary: This article discusses statistical inference for impulse responses and forecast error variance decompositions in sparse, structural high-dimensional vector autoregressive (SVAR) systems. The authors propose consistent estimators and valid inference procedures for these parameters in the high-dimensional setting. The traditional methods, like the delta-method, cannot be applied in this context, so the authors use local projection equations to construct estimators of the moving average parameters and the structural impulse responses. They demonstrate that the distribution of the derived estimators has a Gaussian limit and present a bootstrap procedure for estimating this distribution. The article also provides applications of the inference procedure in constructing confidence intervals and conducting tests. The proposed procedure is illustrated through simulations and an empirical application.
JOURNAL OF ECONOMETRICS
(2023)
Article
Mathematical & Computational Biology
Yanda Lang, Joseph W. McKean, Omer Ozturk
Summary: This paper proposes robust meta-analysis procedures for individual studies reporting a wide range of summary statistics in a two-sample problem. The procedures are compared with conventional meta-analysis based on sample means and variances under different error distributions. Results show that the robust meta-analysis procedures have close coverage probabilities to the nominal confidence level and smaller mean square error (MSE) compared to the non-robust ones under various error distributions. The procedures are then applied to the analysis of platelet count reduction in malaria infected patients in Ghana.
STATISTICS IN MEDICINE
(2023)
Article
Economics
Carlos Aller, Lorenzo Ductor, Daryna Grechyna
Summary: The study identifies GDP per capita, the share of fossil fuels in energy consumption, urbanization, industrialization, democratization, the indirect effects of trade, and political polarization as the robust determinants of CO2 emissions per capita. These determinants all negatively impact the environment except political polarization. Additionally, the determinants of CO2 emissions are found to vary depending on a country's level of income per capita.
Article
Social Sciences, Mathematical Methods
David Roodman, James G. MacKinnon, Morten Orregaard Nielsen, Matthew D. Webb
Article
Economics
Brennan S. Thompson, Matthew D. Webb
ECONOMETRICS JOURNAL
(2019)
Article
Economics
Antoine A. Djogbenou, James G. MacKinnon, Morten Orregaard Nielsen
JOURNAL OF ECONOMETRICS
(2019)
Article
Economics
James G. MacKinnon, Morten Orregaard Nielsen, Matthew D. Webb
Summary: The study examines two cluster-robust variance estimators for regression models with clustering in two dimensions and provides conditions for asymptotically valid inferences based on each. It also introduces several wild bootstrap procedures with conditions for their asymptotic validity for each type of t-statistic. Simulation results suggest that using certain bootstrap procedures with one of the t-statistics generally performs well, and an empirical example shows substantial differences between bootstrap inferences and conventional ones.
JOURNAL OF BUSINESS & ECONOMIC STATISTICS
(2021)
Article
Economics
James G. MacKinnon, Matthew D. Webb
JOURNAL OF ECONOMETRICS
(2020)
Article
Pharmacology & Pharmacy
Matthew M. Loiacono, Christopher B. Nelson, Paul Grootendorst, Matthew D. Webb, Laura Lee Hall, Jeffrey C. Kwong, Nicholas Mitsakakis, Stacy Zulueta, Ayman Chit
Summary: The study showed that implementing behavioral peer comparisons (PC) interventions in community pharmacies can increase SIV uptake, especially among large-format pharmacies, with historically low-performing pharmacies showing the greatest impact.
JOURNAL OF THE AMERICAN PHARMACISTS ASSOCIATION
(2021)
Article
Economics
Matthew D. D. Webb
Summary: Cluster-robust inference is now commonly used in empirical research, and the wild cluster bootstrap is often applied when there are few clusters. However, the conventional bootstrap weights can create ambiguities in inference. This study proposes several modifications to the bootstrap procedure and demonstrates through Monte Carlo simulations that a new 6-point bootstrap weight distribution and kernel density estimation approach can improve the reliability of inference. An empirical example further illustrates the implications of these findings.
CANADIAN JOURNAL OF ECONOMICS-REVUE CANADIENNE D ECONOMIQUE
(2023)
Article
Business, Finance
Roger M. White, Matthew D. Webb
Summary: This paper summarizes and promotes the application of randomization inference in accounting research, discussing its use in both small sample and large sample settings, and providing guidance and sample code for researchers looking to implement randomization inference.
JOURNAL OF FINANCIAL REPORTING
(2021)
Article
Economics
James G. MacKinnon
CANADIAN JOURNAL OF ECONOMICS-REVUE CANADIENNE D ECONOMIQUE
(2019)
Article
Demography
Casey Warman, Matthew D. Webb, Christopher Worswick
JOURNAL OF POPULATION ECONOMICS
(2019)