Replication (version 0.1.0)

llratio.imp: Log likelihood ratio and p-value for data imputed with mice

Description

Computes the log likelihood ratio and p-value for data that is imputed with mice

Usage

llratio.imp(step2step3,imp,model,effectsize=FALSE,s.i,
            sample.cov = NULL, sample.mean = NULL, sample.nobs = NULL,
            group = NULL, cluster = NULL, constraints = "", WLS.V = NULL, NACOV = NULL,
            bayes=FALSE,dp=NULL,nchains=2)

Arguments

step2step3

The likelihood ratio values for simulated data as obtained with the ppc.step2step3 function.

imp

A mids object created with the R-package mice

model

The lavaan model that is to be applied to the data

effectsize

Logic; if TRUE, the constraints concern effectsizes.

s.i

A vector of length p holding with indices for the (pooled) standard deviation parameters with which the effect sizes should be computed

sample.cov

Numeric matrix. A sample variance-covariance matrix. The rownames and/or colnames must contain the observed variable names. For a multiple group analysis, a list with a variance-covariance matrix for each group. Note that if maximum likelihood estimation is used and likelihood="normal", the user provided covariance matrix is internally rescaled by multiplying it with a factor (N-1)/N, to ensure that the covariance matrix has been divided by N. This can be turned off by setting the sample.cov.rescale argument to FALSE.

sample.mean

A sample mean vector. For a multiple group analysis, a list with a mean vector for each group.

sample.nobs

Number of observations if the full data frame is missing and only sample moments are given. For a multiple group analysis, a list or a vector with the number of observations for each group.

group

A variable name in the data frame defining the groups in a multiple group analysis.

cluster

The cluster variable for multilevel data (beta!).

constraints

Additional (in)equality constraints not yet included in the model syntax. See model.syntax for more information. Note that the replication hypothesis should not be specified here!

WLS.V

A user provided weight matrix to be used by estimator "WLS"; if the estimator is "DWLS", only the diagonal of this matrix will be used. For a multiple group analysis, a list with a weight matrix for each group. The elements of the weight matrix should be in the following order (if all data is continuous): first the means (if a meanstructure is involved), then the lower triangular elements of the covariance matrix including the diagonal, ordered column by column. In the categorical case: first the thresholds (including the means for continuous variables), then the slopes (if any), the variances of continuous variables (if any), and finally the lower triangular elements of the correlation/covariance matrix excluding the diagonal, ordered column by column.

NACOV

A user provided matrix containing the elements of (N times) the asymptotic variance-covariance matrix of the sample statistics. For a multiple group analysis, a list with an asymptotic variance-covariance matrix for each group. See the WLS.V argument for information about the order of the elements.

bayes

Logic; if TRUE, a Bayesian estimator is used.

dp

blavaan default prior distributions on different types of parameters, typically the result of a call to dpriors(). See the dpriors() help file for more information.

nchains

A scalar indicating the number of chains to be used in the Bayesian analysis. Default value = 2.

Value

pT

The average parameter table

llratio.i

The likelihood ratio values for each of the imputed datasets

pvals

The prior-predictive p-values for each of the imputed datasets

See Also

llratio.f

Examples

# NOT RUN {
#the following example can be used, but takes >10 seconds

#create data
rnorm2 <- function(n,mean,sd) { mean+sd*scale(rnorm(n)) }

#step 1 input
#create/load data
n.o=30 #sample size original data
y.o <- data.frame(y=rnorm2(n.o,0,1),x=rnorm2(n.o,3,1))
n.r=80 #sample size new data
y.r <- data.frame(y=rnorm2(n.r,0.5,1),x=rnorm2(n.r,3,1))
y.r$y[runif(5,1,n.r)] <- NA #random missing data

#blavaan model
model <- '
y ~ x     #regression
y ~1      #intercept not default in lavaan (but is in blavaan)
'

step1.reg <- ppc.step1(y.o=y.o,model=model,n.r=n.r)

#H0: #reg > est, int = est       =       B1>0.302 & B0= -0.878
pT <- step1.reg$pT #parameter table
int.id <- which(pT$lhs=="y"&pT$op=="~1"&pT$rhs=="") #identify B0
reg.id <- which(pT$lhs=="y"&pT$op=="~"&pT$rhs=="x") #identify B1
hyp <- cbind(pT[c(int.id,reg.id),"plabel"],c("=",">"),c(pT[c(int.id,reg.id),"est"]))
print(hyp)
H0 <- paste(hyp[,1],hyp[,2],hyp[,3],collapse="&")

step23.reg <- ppc.step2step3(step1=step1.reg,y.r=NULL,model=model,H0)

y.r$y[runif(5,1,n.r)] <- NA #random missing data
imp <- mice(y.r,maxit=10,m=10)
llratio.imp(step2step3=step23.reg,imp=imp,model=model)
# }