| mxBootstrapStdizeRAMpaths {OpenMx} | R Documentation |
Uses the bootstrap distribution of a RAM model's free parameters to produce a bootstrap distribution of standardized path coefficients. Model must have already been run through mxBootstrap().
mxBootstrapStdizeRAMpaths(model, bq=c(.25,.75),
method=c('bcbci','quantile'), returnRaw=FALSE)
model |
An MxModel that uses RAM expectation and has already been run through |
bq |
Quantiles corresponding to the lower and upper limits, respectively, of the bootstrap confidence interval. |
method |
Character string. One of 'bcbci' or 'quantile'. |
returnRaw |
Logical; should the function return the raw bootstrapping results? Defaults to |
In effect, what mxBootstrapStdizeRAMpaths() does is use the point estimates from each bootstrap replication to produce mxStandardizeRAMpaths() output for each replication. The output of mxStandardizeRAMpaths() has one entry for each nonzero path coefficient. Therefore, it is possible (though unlikely) that the number of nonzero paths, or which elements of the A and S RAM matrices are nonzero, may vary among bootstrap replications. Such an occurrence defies simple summary of the standardized paths' bootstrapping results. In this case, a raw list of bootstrapping results is returned, with a warning, if returnRaw=TRUE, or an error is throw if returnRaw=FALSE.
mxBootstrapStdizeRAMpaths() ignores any submodels of model. That is, it must be able to find, in the MxModel it is provided, a RAM expectation and an MxComputeBootstrap object. It can be run on submodels directly,
Under extraordinary circumstances described above, mxBootstrapStdizeRAMpaths() may return a list object. In ordinary circumstances, if returnRaw=FALSE (default), it returns a dataframe containing, inter alia, the standardized path coefficients as estimated from the real data, their bootstrap SEs, and the lower and upper limits of a bootstrap confidence interval. If instead returnRaw=TRUE, mxBootstrapStdizeRAMpaths() returns a matrix containing the raw bootstrap results; this matrix has one column per nonzero path coefficient, and one row for each succesfully converged bootstrap replication.
mxBootstrap(), mxStandardizeRAMpaths(), mxBootstrapEval, mxSummary
require(OpenMx)
data(myFADataRaw)
myFADataRaw <- myFADataRaw[,c("x1","x2","x3","x4","x5","x6")]
dataRaw <- mxData( observed=myFADataRaw, type="raw" )
resVars <- mxPath( from=c("x1","x2","x3","x4","x5","x6"), arrows=2,
free=TRUE, values=c(1,1,1,1,1,1),
labels=c("e1","e2","e3","e4","e5","e6") )
latVar <- mxPath( from="F1", arrows=2,
free=TRUE, values=1, labels ="varF1" )
facLoads <- mxPath( from="F1", to=c("x1","x2","x3","x4","x5","x6"), arrows=1,
free=c(FALSE,TRUE,TRUE,TRUE,TRUE,TRUE), values=c(1,1,1,1,1,1),
labels =c("l1","l2","l3","l4","l5","l6") )
means <- mxPath( from="one", to=c("x1","x2","x3","x4","x5","x6","F1"), arrows=1,
free=c(TRUE,TRUE,TRUE,TRUE,TRUE,TRUE,FALSE),
values=c(1,1,1,1,1,1,0),
labels =c("meanx1","meanx2","meanx3",
"meanx4","meanx5","meanx6",NA) )
oneFactorModel <- mxModel("Common Factor Model Path Specification", type="RAM",
manifestVars=c("x1","x2","x3","x4","x5","x6"), latentVars="F1",
dataRaw, resVars, latVar, facLoads, means)
oneFactorFit <- mxRun(oneFactorModel)
set.seed(170505)
## Not run: oneFactorBoot <- mxBootstrap(oneFactorFit)
mxBootstrapStdizeRAMpaths(oneFactorBoot)
## End(Not run)