Did changes to how qgraph in R calculate centrality measures when measuring the network using EBICglasso algorithm (using cor_auto function) occur between Dec 2022 and June 2023?
I am running the same script with the same data, but getting slightly different result after updating the R and RStudio lately, and for the life of me cannot see how and why the differences occur. The differences are slight, within error ranges, but the scale for expected influence and other centrality measures has changed (from [-2;1] to [0;1]) and some of the variables now contain missing values. Any advice/info on changes in calculations?
Cannot find anything on their package info or seem to fail to understand.
In version 1.8 I changed the default of the scale argument, which was previously using z-values but that was a weird choice of default. You can see the package version changes here.