JB
Joe Bak-Coleman
@jbakcoleman.bsky.social
Associate Research Scientist at Columbia journalism. Harvard BKC affiliate. Comp. soc. science, collective behavior, stats
2.2k followers681 following1k posts
I think the response letter is a bit misleading because it analyzes distinct data on a distinct timescale ignoring the earlier months and asks a different question. It’s also frustrating because meta has the data to answer the concern.
The studies issues with causality are (imo) a much bigger issue for the estimates but this is a problem even if you believe withdrawal and exposure are symmetric and spillover effects don’t matter.
JB
Joe Bak-Coleman
@jbakcoleman.bsky.social
Associate Research Scientist at Columbia journalism. Harvard BKC affiliate. Comp. soc. science, collective behavior, stats
2.2k followers681 following1k posts