Generate wild bootstrap replicates of a statistic for a linear mixed-effects model.
# S3 method for lmerMod
wild_bootstrap(
model,
.f,
B,
hccme = c("hc2", "hc3"),
aux.dist = c("mammen", "rademacher", "norm", "webb", "gamma"),
.refit = TRUE
)# S3 method for lme
wild_bootstrap(
model,
.f,
B,
hccme = c("hc2", "hc3"),
aux.dist = c("mammen", "rademacher", "norm", "webb", "gamma"),
.refit = TRUE
)
wild_bootstrap(model, .f, B, hccme, aux.dist, .refit = TRUE)
The returned value is an object of class "lmeresamp".
The model object you wish to bootstrap.
A function returning the statistic(s) of interest.
The number of bootstrap resamples.
either "hc2"
or "hc3"
, indicating which
heteroscedasticity consistent covariance matrix estimator to use.
one of "mammen"
, "rademacher"
, "norm"
,
"webb"
, or "gamma"
indicating which auxiliary
distribution to draw the errors from
a logical value indicating whether the model should be refit to
the bootstrap resample, or if the simulated bootstrap resample should be
returned. Defaults to TRUE
.
The wild bootstrap algorithm for LMEs implemented here was outlined by Modugno & Giannerini (2015). The algorithm is outlined below:
Draw a random sample equal to the number of groups (clusters) from an auxillary distribution with mean zero and unit variance. Denote these as \(w_1, \ldots, w_g\).
Calculate the selected heteroscedasticity consistent matrix estimator for the marginal residuals, \(\tilde{v}_i\)
Generate bootstrap responses using the fitted equation: \(y_i^* = X_i \beta + \tilde{v}_i w_j\)
Refit the model and extract the statistic(s) of interest.
Repeat steps 2-4 B times.
Modugno, L., & Giannerini, S. (2015). The Wild Bootstrap for Multilevel Models. Communications in Statistics -- Theory and Methods, 44(22), 4812--4825.
Examples are given in bootstrap
parametric_bootstrap
, resid_bootstrap
,
case_bootstrap
, reb_bootstrap
,
wild_bootstrap
for more details on a specific bootstrap.
bootMer
in the lme4 package for an
implementation of (semi-)parametric bootstrap for mixed models.