Ancestors

Toot

Written by Gelman et al. on 2025-01-23 at 16:12

When estimating a treatment effect with a cluster design, you need to include varying slopes, even if the fit gives warning messages. https://statmodeling.stat.columbia.edu/2025/01/23/slopes/

=> More informations about this toot | More toots from statmodeling_bot@bayes.club

Descendants

Written by Leo Ré Jorge on 2025-01-24 at 08:28

@statmodeling_bot I was pretty shocked by this one... I can't recall many (if any) papers in ecology including varying slopes in this kind of blocked design when there is a single observation per treatment/block combination. I myself probably have a dozen papers published with this error

=> More informations about this toot | More toots from LeoRJorge@mastodon.social