Calculating Mcdonald's Omega After Multiple Imputation

[Nick]

Hello!

I have a dataset which underwent a multiple imputation procedure using the mice package in R, resulting in 5 plausible datasets. I am trying to make a simple descriptives table including Mcdonald's Omega values for several scales. My understanding (correct me if I'm wrong!) is that there is no prebuilt function to do this for multiply imputed datasets.

One workaround solution I'm considering is to get the omega from a CFA of each scale using the lavaan.mi package. See below: 

mids <- read_rds("imputeddatasets.rds")

imp_list <- complete(mids, action = "all")

mod2 <- '
Factor1 =~ item_1 + item_2 + item_3 +  
                  item_4 + item_5 + item_6 
'

fit2 <- cfa.mi(
  model = mod2,
  data = imp_list,
  estimator = "DWLS",
  std.lv = TRUE,
  ordered = TRUE
)
summary(fit2, fit.measures = TRUE, standardized = TRUE)
semTools::reliability(fit2)

I guess my question boils down to: Would this method be sufficient and "close enough" for a descriptives table in a manuscript? Or is there a better way to accomplish what I'm looking for? Thanks for any help anyone can provide here!

[Terrence Jorgensen]

semTools::reliability() has been deprecated for years.  It has been superseded by the more comprehensive function semTools::compRelSEM(),  which does accept a lavaan.mi object (using the lavaan.mi package) as easily as a lavaan object.

Best,

Terrence D. Jorgensen    (he, him, his)
Assistant Professor, Methods and Statistics
Research Institute for Child Development and Education, the University of Amsterdam
http://www.uva.nl/profile/t.d.jorgensen