Error using compRelSEM with lavaan.mi object

81 views

Skip to first unread message

Stephanie Kliethermes

unread,

Sep 14, 2023, 11:32:29 AM9/14/23

to lavaan

Hi lavaan group - I’ve been working my way through some psychometric analyses of a new tool we are developing in the sports sciences and I’ve been spinning my wheels a bit on an error I get when trying to obtain indicator reliability on our 3-factor CFA model using lavaan and multiply imputed datasets. I would like to rely on the compRelSEM function in SEMtools but am struggling to get output from my lavaan.mi object. As far as I can tell, my error occurs when trying to pull the variances from the @phiList. When looking at the @phiList, it appears as though the covariance matrices are accurate for all imputations. The error is as follows:

Error in PHI[[block.label[b]]] <- Reduce("+", lapply(phiList, "[[", i = b))/m :

attempt to select less than one element in OneIndex

If I run each of the imputed datasets separately through compRelSEM I can get estimates. But I'm struggling with the pooled estimates. Any help or guidance would be appreciated.

# Define the CFA model

model3a <- '

# Factor indicators

Factor1 =~ social1 + social2 + social3 +social4 + social5 + social6

Factor2 =~ motivation1 + motivation3 + motivation10+ motivation11+ motivation12

Factor3 =~ physical1 + physical4+physical5

# Factor covariances

Factor1 ~~ Factor2

Factor1 ~~ Factor3

Factor2 ~~ Factor3

# run lavaan and imputation in one step

fittedmodel <- runMI(model3a,

data=data,

m = 5,

miPackage="mice",

fun="cfa",

estimator="DWLS",

se="robust",

test="scaled.shifted",

meanstructure = TRUE)

#summary

summary(fittedmodel, fit.measures=TRUE, standardized=TRUE)

#Indicator Reliability

compRelSEM(fittedmodel, tau.eq=FALSE)

Error in PHI[[block.label[b]]] <- Reduce("+", lapply(phiList, "[[", i = b))/m :

attempt to select less than one element in OneIndex

Terrence Jorgensen

unread,

Sep 15, 2023, 7:49:05 AM9/15/23

to lavaan

I don't have access to your data, but it works fine on the models from the ?runMI help-page examples. The default is already tau.eq=FALSE, but I don't see what could be causing this problem. Do you have the latest software installed? It is probably a bug that I already fixed.

remotes::install_github("simsem/semTools/semTools")

https://github.com/simsem/semTools/wiki

Terrence D. Jorgensen

Assistant Professor, Methods and Statistics

Research Institute for Child Development and Education, the University of Amsterdam