Full Vignette for Revision #37
Description
Regarding the models, I think the paper needs more clarity on the specifics, especially around the pooled regression approach. I think some sort of visualization of the selection process for the first-born full siblings--> standard OLS regression in a wide data set--> specifics on the difference scores and the pooled regression (and use of the sibling specific data to test interact) will help. The good news is that this is all pretty straightforward and non-technical, so it can be easily explained. This explanation interacts with the details of the [discord] package. I did not find the Git repository content or even the [discord] documentation to be especially accessible. The Supplementary Material in the paper (illustrating the workflow from for the test of neuroticism to mental health) is good, but I would encourage adding more information-- e.g., on the entire workflow. Show the OLS lm. Yes, most people can do this easily enough, but also show how you select the first-born sibling, then show the results of the lm-- basically, illustrate the full workflow. There's a "voila feeling" to the Supp Material that I think could be deconstructed to make this document a more complete companion to this paper.
My advice: include everything that is needed here to do end-to-end. We don't need NLSY details of course b/c this will be different for all data sets, but is the starting data set wide or long? How do you do the first-born selection? How does the discord package order the siblings for the pooled regression? Work toward making this paper a complete, stand-alone, everything-you-need-to-get-it-done paper. I understand its not a Tutorial per se, but it's somewhere close!