Digitale et al 2022. Tutorial on DAGs. doi.org/10.1016/j.jc....doi.org/10.1001/jama...doi.org/10.1098/rspb...doi.org/10.1093/ije/...
If you didn't know, now you know: In general causal graphs, X>Y by default means "X may or may not cause Y," which means "I don't know" (in classical logic), but in *acyclic* graphs an arrow is a *stronger* assumption than an absence of an edge: if X>Y then not Y>X but not Y>X says nothing about X>Y
💯 Especially if "methodological" is understood as including logic / definitions+argument structure.
If you want to find out whether something has been replicated, you can also go to forrt-replications.shinyapps.io/fred_explorer/, choose all studies, and enter a search term. We try to keep track of all replication attempts.
academic papers should include an "Author's Cut" button that easily toggles between the paper they wanted published, and the version with all the stuff forced on them to get it through review
BTW, this is also what you get when you use the HMeta-d' package to obtain an m-ratio estimate. The denominator that this package uses for the m-ratio index is a (basic) *point* estimate of d'. There's no way to fix the m-ratio as the "theory" is lacking, but it *can* be slightly less ridiculous 🕊️
Of course everyone in the Bsky community knows all these results off the top of their head. But in case you have colleagues or students who may not I thought I'd share. www.sciencedirect.com/science/arti...
Drafts of the first chapters of my brms Book: Applied Bayesian Regression Modelling Using R and Stan are online: paulbuerkner.com/software/brm... Check it out and let me know what you think!
Check out his papers on SEM vs marginal structural models (www.tandfonline.com/doi/full/10....www.tandfonline.com/doi/full/10....)
The use of structural equation models for causal inference from panel data is critiqued in the causal inference literature for unnecessarily relying on a large number of parametric assumptions, and...