Cluster robust wild bootstrap for linear models {Rfast2}

Cluster robust wild bootstrap for linear models

Description

Cluster robust wild bootstrap for linear models.

Usage

wild.boot(y, x, cluster, ind = NULL, R = 999, parallel = FALSE)

Arguments

y

The dependent variable, a numerical vector with numbers.

x

A matrix or a data.frame with the indendent variables.

cluster

A vector indicating the clusters.

ind

A vector with the indices of the variables for which wild bootstrap p-values will be computed. If NULL (default value), the p-values are computed for each variable.

R

The number of bootstrap replicates to perform.

parallel

Do you want the process to take place in parallel? If yes, then set this equal to TRUE.

Details

A linear regression model for clustered data is fitted. For more information see Chapter 4.21 of Hansen (2019).

Value

A matrix with 5 columns, the estimated coefficients of the linear model, their cluster robust standard error, their cluster robust test statistic, their cluster robust p-value, and their cluster robust wild bootstrap p-value.

Author(s)

Michail Tsagris and Stefanos Fafalios.

R implementation and documentation: Michail Tsagris [e-mail] and Stefanos Fafalios [e-mail].

References

Cameron A. Colin, Gelbach J.B., and Miller D.L. (2008). Bootstrap-Based Improvements for Inference with Clustered Errors. The Review of Economics and Statistics 90(3): 414-427.

See Also

gee.reg, cluster.lm

Examples

y <- rnorm(200)
id <- sample(1:20, 200, replace = TRUE)
x <- matrix( rnorm(200 * 3), ncol = 3 )
wild.boot(y, x, cluster = id)