After re-reading the Hallgren et al. paper and the web note, I think for scatter plots, we should
a. Default to using factor scores (or observed variables) from the input data
b. Given the confusion about which trend line to add (e.g., model-implied or data-driven), I think it's best not to add a trend line.
c. Create a function to compute casewise residuals (see semdplot).
https://www.statmodel.com/examples/webnotes/webnote%2020.pdf
After re-reading the Hallgren et al. paper and the web note, I think for scatter plots, we should
a. Default to using factor scores (or observed variables) from the input data b. Given the confusion about which trend line to add (e.g., model-implied or data-driven), I think it's best not to add a trend line. c. Create a function to compute casewise residuals (see
semdplot).Originally posted by @marklhc in https://github.com/Gengrui-Zhang/R2spa/issues/23#issuecomment-1582960068