R: Bootstrapping multilevel mediation model (without boot...

boot.modmed.mlm.custom {multilevelmediation}

R Documentation

Bootstrapping multilevel mediation model (without boot package)

Description

Bootstrapping multilevel mediation model (without boot package)

Usage

boot.modmed.mlm.custom(
  data,
  L2ID,
  ...,
  return.type = "all",
  modval1 = NULL,
  modval2 = NULL,
  nrep = 500,
  boot.type = c("caseboth", "case2", "case1", "resid"),
  parallel.type = c("lapply", "parallel", "furrr"),
  ncores = NULL,
  seed = NULL
)

Arguments

`data`	Data frame in long format. The function will do restructuring using `stack_bpg`.
`L2ID`	Name of column that contains grouping variable in 'data' (e.g., "SubjectID")
`...`	Arguments passed to `modmed.mlm` or `lme` to define the mediation analysis model or do estimation, respectively.
`return.type`	Character that defines what information to extract from the model. Default and options are in `extract.modmed.mlm`. As examples, "indirect" will compute the indirect effect, "all" will save all random and fixed effects for possible additional computations, "indirect.diff" will compute the difference in the indirect effect at two values of a possible moderating variable.
`modval1`	(Optional) Numeric. If the model has a moderator, this value will be passed to `extract.modmed.mlm` to compute the indirect effect or other effects at that value. See `extract.modmed.mlm` for details.
`modval2`	(Optional). If the model has a moderator, it is possible to compute the difference in the indirect at two values of the moderator. If given and an appropriate option for such a difference is chosen for `type`, this value and that of `modval1` will be passed to `extract.modmed.mlm` to compute and save the difference. This is useful for obtaining a CI for the difference in the indirect effect at two different levels of the moderator.
`nrep`	Number of bootstrap replications to perform. Pick a small number just for testing purposes, something larger (e.g., 1000 or more) for analyses.
`boot.type`	Character indicating the type of bootstrapping to perform. Options are: "caseboth", "case2", "case1", or "resid".
`parallel.type`	Character indicating type of parallel processing (if any) to use. Options are "lapply" (no parallel processing),"parallel" (uses `parallel` package), or "furrr" (uses `furrr` package.
`ncores`	Integer indicating the number of processing cores to attempt to use.
`seed`	Integer to set random number seed, for replication purposes. Note that replication may be somewhat R version or platform dependent.

Details

This function was written to do all four kinds of bootstrapping outlined in Falk, Vogel, Hammami & Miočević (in press): case resampling at both levels, at level 2 only, at level 1 only, and the residual-based bootstrap (e.g., see Hox and van de Schoot, 2013; van der Leeden, Meijer, & Busing, 2008). These functions also support moderated mediation. See also modmed.mlm. Note that nlm was used as the optimizer for some of the examples below as it was found to be faster for the models/simulations studied by Falk et al (in press). Note that Level 1 only bootstrapping is typically not recommended. See Falk et al. (in press) for details.

This function is different from the original functions used for the publication and that as of this writing still appear here: boot.modmed.mlm and here: bootresid.modmed.mlm . The present function seeks to unify case bootstrapping and residual-based bootstrapping in the same function. Furthermore, this newer function is also aimed at attempting to bypass the need for using the boot package to do computations and parallel processing. Some performance gains in terms of speed have been observed via use of this function instead of boot in conjunction with boot.modmed.mlm. Although somewhat slower, furrr can also be used if one would like a progress bar.

Value

A list with the following elements. Note that t0 and t are intended to trick the boot package into working with some if its functions.

call Call/arguments used when invoking this function. Useful for later extracting things like indirect effect.
t0 Parameter estimates based on the dataset.
t Bootstrap distribution of all parameter estimates.
model Fitted model to restructured data as one would obtain from modmed.mlm.
conv Whether model fit to restructured dataset converged.
args Arguments used when calling modmed.mlm. Useful for later extracting things like indirect effect.

References

Bauer, D. J., Preacher, K. J., & Gil, K. M. (2006). Conceptualizing and testing random indirect effects and moderated mediation in multilevel models: New procedures and recommendations. Psychological Methods, 11(2), 142–163. doi:10.1037/1082-989X.11.2.142

Falk, C. F., Vogel, T., Hammami, S., & Miočević, M. (in press). Multilevel mediation analysis in R: A comparison of bootstrap and Bayesian approaches. Behavior Research Methods. doi:10.3758/s13428-023-02079-4 Preprint: doi:10.31234/osf.io/ync34

Hox, J., & van de Schoot, R. (2013). Robust methods for multilevel analysis. In M. A. Scott, J. S. Simonoff & B. D. Marx (Eds.), The SAGE Handbook of Multilevel Modeling (pp. 387-402). SAGE Publications Ltd. doi:10.4135/9781446247600.n22

van der Leeden, R., Meijer, E., & Busing, F. M. T. A. (2008). Resampling multilevel models. In J. de Leeuw & E. Meijer (Eds.), Handbook of Multilevel Analysis (pp. 401-433). Springer.

Examples




data(BPG06dat)

# Note that for all examples below, nrep should be increased to something
#  MUCH larger (e.g., 1000). Small values here are used only so that the code
#  runs relatively quickly when tested.

# double bootstrap, no parallel processing
boot.result<-boot.modmed.mlm.custom(BPG06dat, nrep=10, L2ID="id", X="x", Y="y", M="m",
  boot.type="caseboth",
  control=list(opt="nlm"), seed=1234)

extract.boot.modmed.mlm(boot.result, type="indirect", ci.conf=.95)

# residual bootstrap, parallel package
boot.result<-boot.modmed.mlm.custom(BPG06dat, nrep=10, L2ID="id", X="x", Y="y", M="m",
  boot.type="resid", random.a=TRUE, random.b=TRUE,
  parallel.type="parallel",ncores=2,seed=2299,
  control=list(opt="nlm"))

extract.boot.modmed.mlm(boot.result, type="indirect", ci.conf=.95)



# Example with moderation
data(simdat)

# moderation
boot.result<-boot.modmed.mlm.custom(simdat, nrep=5, L2ID = "L2id", X = "X", Y = "Y", M = "M",
   boot.type="caseboth",
   random.a=TRUE, random.b=TRUE, random.cprime=TRUE,
   moderator = "mod", mod.a=TRUE, mod.b=TRUE,
   random.mod.a = TRUE, random.mod.b = TRUE,
  parallel.type="parallel",ncores=2,seed=2299)

extract.boot.modmed.mlm(boot.result, type="indirect")

# indirect effect point estimate and 95% CI when moderator = 0
extract.boot.modmed.mlm(boot.result, type="indirect", modval1=0)

# indirect effect point estimate and 95% CI when moderator = 1
 extract.boot.modmed.mlm(boot.result, type="indirect", modval1=1)

# indirect effect difference point estimate and 95% CI
extract.boot.modmed.mlm(boot.result, type="indirect.diff",
  modval1=0, modval2=1)

[Package multilevelmediation version 0.3.1 Index]