My latest WP, a restructured version of the piece in the quoted thread. Helpful for thinking clearly about social media RCTs and their limits, useful for teaching about spillover in causal inference SocArxiv: osf.io/preprints/so... Dropbox: www.dropbox.com/scl/fi/zj43w... (1/5)

Is social media to blame for teens’ mental health decline? Intuition says yes—but RCTs find only small short-term effects when users quit. This new preprint argues these studies don’t prove social media isn’t to blame. Here’s why: 3 reasons.

Trends in emergency department visits overstate actual deterioration in child mental health significantly due to changes in screening, definition, and coding of mental illness, from Han Choi, Adriana Corredor-Waldron, Janet Currie, and @cfelton.bsky.social https://www.nber.org/papers/w33550

Instrumental variables guide: journals.sagepub.com/doi/full/10.... '13 Reasons Why' and self-harm: sociologicalscience.com/articles-v10...

Something I only just noticed, but can no longer unsee, is that the title of Imbens & Rubin contains a usage error. Either "Statistics" should be changed to "Statistical," followed by a comma, or it should read "[...] for Statistics and the Social and Biomedical Sciences."

From the data-is-plural.com newsletter (which you should subscribe to). Fun, open-access dataset that you can use for stats problem sets. erictleung.com/pixarfilms/