likRatioFit: Find the likelihood ratio (or Bayes factor) based on the bivariate distribution of fit indices

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)

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.

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.

Author

Sunthud Pornprasertmanit (psunthud@gmail.com)

Examples

Run this code

if (FALSE) {
# 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)
}

Run the code above in your browser using DataLab

The Learning Leader's Guide to AI Literacy