A few years ago, I completely rewrote the runMI() functionality in semTools, capitalizing on lavaan's lavaanList() function and writing dedicated methods for the lavaan.mi-class that inherits from lavaanList-class. There have been some odd problems with R finding those methods, even when users properly load the library(semTools) instead of inappropriately relying on lazy loading (e.g., running semTools::runMI() without loading semTools).
I am deprecating the runMI() functionality from the next version semTools, moving all that functionality to the new lavaan.mi package, which fixes many bugs, implements many new features, and better synchronizes the user experience between lavaan with (in)complete data and lavaan.mi for multiple imputations.
PRIOR TO USING THE NEW PACKAGE, please closely read the README and NEWS files to learn about small changes to the user interface (e.g., some updated argument names, how summary() works).
If you have the time to volunteer by re-running your previous runMI() analysis using the new lavaan.mi package (which does NOT have a runMI() function, only the lavaan.mi() function), please install the latest software from GitHub (using the remotes package for installation):
remotes::install_github("yrosseel/lavaan")
remotes::install_github("TDJorgensen/lavaan.mi")
remotes::install_github("simsem/semTools/semTools")
Thank you for your help, and please be patient as I find time to address any issues.
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