I like this answer, sadly I have p>20! potential IVs and p>>>N (obviously). I'm thinking Belloni et al (2012) won't exactly work (computationally) for such huge p, but I can still use their model comparison criteria if I can partition my IVs into m<<p 1st stage models? Does this sound plausible? - ThreadSky

clintent-to-treat.bsky.social • 135 days ago

I like this answer, sadly I have p>20! potential IVs and p>>>N (obviously).

I'm thinking Belloni et al (2012) won't exactly work (computationally) for such huge p, but I can still use their model comparison criteria if I can partition my IVs into m<

Does this sound plausible?

Comments

Posting Rules

Comments

Posting Rules

Reply