Bootstrap tests: how many bootstraps?
- 1 January 2000
- journal article
- research article
- Published by Taylor & Francis in Econometric Reviews
- Vol. 19 (1) , 55-68
- https://doi.org/10.1080/07474930008800459
Abstract
In practice, bootstrap tests must use a finite number of bootstrap samples. This means that the outcome of the test will depend on the sequence of random numbers used to generate the bootstrap samples, and it necessarily results in some loss of power. We examine the extent of this power loss and propose a simple pretest procedure for choosing the number of bootstrap samples so as to minimize experimental randomness. Simulation experiments suggest that this procedure will work very well in practice.Keywords
All Related Versions
This publication has 11 references indexed in Scilit:
- Exact Inference Methods for First-Order Autoregressive Distributed Lag ModelsEconometrica, 1998
- Graphical Methods for Investigating the Size and Power of Hypothesis TestsThe Manchester School, 1998
- Bootstrap Critical Values for Tests Based on Generalized-Method-of-Moments EstimatorsEconometrica, 1996
- Bootstrapping time series modelsEconometric Reviews, 1996
- Bootstrap-based critical values for the information matrix testJournal of Econometrics, 1994
- Prepivoting Test Statistics: A Bootstrap View of Asymptotic RefinementsJournal of the American Statistical Association, 1988
- Finite Sample Properties and Asymptotic Efficiency of Monte Carlo TestsThe Annals of Statistics, 1986
- Barnard's Monte Carlo Tests: How Many Simulations?Journal of the Royal Statistical Society Series C: Applied Statistics, 1979
- A Simplified Monte Carlo Significance Test ProcedureJournal of the Royal Statistical Society Series B: Statistical Methodology, 1968
- Modified Randomization Tests for Nonparametric HypothesesThe Annals of Mathematical Statistics, 1957