likRatioFit: Find the likelihood ratio (or Bayes factor) based on the...

In simsem: SIMulated Structural Equation Modeling

Description

Find the log-likelihood of the observed fit indices on Model 1 and 2 from the real data on the bivariate sampling distribution of fit indices fitting Model 1 and Model 2 by the datasets from the Model 1 and Model 2. Then, the likelihood ratio is computed (which may be interpreted as posterior odd). If the prior odd is 1 (by default), the likelihood ratio is equivalent to Bayes Factor.

Usage

likRatioFit(outMod1, outMod2, dat1Mod1, dat1Mod2, dat2Mod1, dat2Mod2, 
usedFit=NULL, prior=1)

Arguments

outMod1	lavaan that saves the analysis result of the first model from the target dataset
outMod2	lavaan that saves the analysis result of the second model from the target dataset
dat1Mod1	SimResult that saves the simulation of analyzing Model 1 by datasets created from Model 1
dat1Mod2	SimResult that saves the simulation of analyzing Model 2 by datasets created from Model 1
dat2Mod1	SimResult that saves the simulation of analyzing Model 1 by datasets created from Model 2
dat2Mod2	SimResult that saves the simulation of analyzing Model 2 by datasets created from Model 2
usedFit	Vector of names of fit indices that researchers wish to getCutoffs from. The default is to getCutoffs of all fit indices.
prior	The prior odds. The prior probability that Model 1 is correct over the prior probability that Model 2 is correct.

Value

The likelihood ratio (Bayes Factor) in preference of Model 1 to Model 2. If the value is greater than 1, Model 1 is preferred. If the value is less than 1, Model 2 is preferred.

Author(s)

Sunthud Pornprasertmanit ([e-mail])

See Also

SimResult for a detail of simResult pValueNested for a nested model comparison by the difference in fit indices pValueNonNested for a nonnested model comparison by the difference in fit indices

Examples

## Not run: 
# Model A; Factor 1 --> Factor 2; Factor 2 --> Factor 3
library(lavaan)
loading <- matrix(0, 11, 3)
loading[1:3, 1] <- NA
loading[4:7, 2] <- NA
loading[8:11, 3] <- NA
path.A <- matrix(0, 3, 3)
path.A[2, 1] <- NA
path.A[3, 2] <- NA
model.A <- estmodel(LY=loading, BE=path.A, modelType="SEM", indLab=c(paste("x", 1:3, sep=""), 
	paste("y", 1:8, sep="")))

out.A <- analyze(model.A, PoliticalDemocracy)

# Model A; Factor 1 --> Factor 3; Factor 3 --> Factor 2
path.B <- matrix(0, 3, 3)
path.B[3, 1] <- NA
path.B[2, 3] <- NA
model.B <- estmodel(LY=loading, BE=path.B, modelType="SEM", indLab=c(paste("x", 1:3, sep=""), 
	paste("y", 1:8, sep="")))

out.B <- analyze(model.B, PoliticalDemocracy)

loading.mis <- matrix("runif(1, -0.2, 0.2)", 11, 3)
loading.mis[is.na(loading)] <- 0

# Create SimSem object for data generation and data analysis template
datamodel.A <- model.lavaan(out.A, std=TRUE, LY=loading.mis)
datamodel.B <- model.lavaan(out.B, std=TRUE, LY=loading.mis)

# Get sample size
n <- nrow(PoliticalDemocracy)

# The actual number of replications should be greater than 20.
output.A.A <- sim(20, n=n, model.A, generate=datamodel.A) 
output.A.B <- sim(20, n=n, model.B, generate=datamodel.A)
output.B.A <- sim(20, n=n, model.A, generate=datamodel.B)
output.B.B <- sim(20, n=n, model.B, generate=datamodel.B)

# Find the likelihood ratio ;The output may contain some warnings here. 
# When the number of replications increases (e.g., 1000), the warnings should disappear.
likRatioFit(out.A, out.B, output.A.A, output.A.B, output.B.A, output.B.B)

## End(Not run)

simsem documentation built on March 29, 2021, 1:07 a.m.