statstas.datascience.blue
Social data scientist
AMA about survey statistics, sampling weights, small area estimation, multilevel models, latent class analysis, structural equation models, #rstats, #Stata.
Scraping contents of this account is not permitted.
1,253 posts
2,298 followers
1,306 following
Prolific Poster
comment in response to
post
Yes, it's a QR decomposition, a bespoke one modified by Ross from LINPACK fortran code.
The key difference from LINPACK (in src/appl/dqrdc2.f) is that it pivots only when really needed, so that computing sequential "this beta adjusted for the previous ones" is more straightforward.
comment in response to
post
@tyk314.net may have some aging research opportunities for you.
comment in response to
post
@ologies.bsky.social you had an episode not too long ago
comment in response to
post
1. I didn’t say “ban cars”
2. I didn’t say “ban on-street parking.”
3. Too many cars & on-street parking make traffic congestion worse. Those are facts.
4. Straw man arguments suck.
5. Populist politicians who claim to want to “make traffic better” almost always attack the wrong things, on purpose.
comment in response to
post
Methodologists have already been hit with reduced access to the data, duuh
comment in response to
post
Maybe you should identify the cohort effect by bringing immigrants. They would have the same age, and by getting embedded to the society will eventually consume the same period effects. But their cohort effects are those of the country of origin (which of course may have even greater variance).
comment in response to
post
A way to think about the Box 2 model is that while A + P - C == 0 in actual world, the random effect mechanics provides shrinkage and the rhs is not exactly zero, showing apparent something (you can’t produce identification out of nothing, shrinkage is as bad a way as any other like nonlinearity).
comment in response to
post
Maybe “Captain’s Daughter” by Alexander Pushkin is a decent substitute.