Vuong Test with Lavaan Object

[drsco...@gmail.com]

Good morning,

Is anyone familiar with the package nonnest2? I'm trying to compare two sem() models using the vuongtest(fit.1, fit.2) and am getting the following error. Does anyone know what that means?

> vuongtest(fit.1, fit.2)

Error in Score.mat[wi, ] <- -scores.H1 %*% Delta :
  number of items to replace is not a multiple of replacement length

CODE:

model.1 <-  ' Y ~ x1 + x2 + x3 + x4 + x5 + x6'

model.2 <-  ' Y ~ x1 + x2 + x3 + x4 + x5 + x6 + x7'

fit.1 <- sem(model.1, df, missing = "fiml")

fit.2 <- sem(model.2, df, missing = "fiml")

[Mauricio Garnier-Villarreal]

Both models have to have the same items. Cant test adding x7

could do

model.1 <-  ' Y ~ x1 + x2 + x3 + x4 + x5 + x6 + 0*x7 '

model.2 <-  ' Y ~ x1 + x2 + x3 + x4 + x5 + x6 + x7'

[drsco...@gmail.com]

Good morning.

I tried 

model.1 <-  ' Y ~ x1 + x2 + x3 + x4 + x5 + x6 + 0*x7 '

model.2 <-  ' Y ~ x1 + x2 + x3 + x4 + x5 + x6 + x7'

but am still getting the error

Error in Score.mat[wi, ] <- -scores.H1 %*% Delta :
  number of items to replace is not a multiple of replacement length

[Christian Arnold]

The problem is probably related to missing data:

library(lavaan)

library(nonnest2)

# Simulate data

m.pop <- "Y ~ 1 * x1 + 0.9 * x2 + 0.9 * x3 + 0.8 * x4 + 0.9 * x5 + 0.8 * x6 + 0.9 * x7"

set.seed(1)

df <- simulateData(m.pop, sample.nobs = 1000)

m1 <- "Y ~ x1 + x2 + x3 + x4 + x5 + x6"

m2 <- "Y ~ x1 + x2 + x3 + x4 + x5 + x6 + x7"

fit1 <- sem(m1, df)

fit2 <- sem(m2, df)

# Run Vuong test ---- works fine

vuongtest(fit1, fit2)

# Quick and dirty delete some data

deleteData <- function(df, missings, rows, cols) {

  del <- list(

    rows = floor(runif(missings, min = 1, max = rows)),

    cols = floor(runif(missings, min = 1, max = cols))

  )

  for(m in 1 : missings) df[del$rows[m], del$cols[m]] <- NA

  return(df)

}

df <- deleteData(df, 200, nrow(df), ncol(df))

fit1 <- sem(m1, df, missing = "fiml")

fit2 <- sem(m2, df, missing = "fiml")

# Run xuong test ---- Error in Score.mat[wi, ]

vuongtest(fit1, fit2)

As you can see, the Vuong test works fine without missing data (apart from the warning message). You could try to delete all rows with missing data from your data frame.

---

--
You received this message because you are subscribed to the Google Groups "lavaan" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lavaan+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/lavaan/1585a0d5-1ff5-4881-93c8-e19f9f217bc4n%40googlegroups.com.