Quant III, Homework 1
a) get the data for the huber et al study into R
b) run some individual country logits (you can
just do a few countries)
c) take the full data set and run with
interactions (so the mlevel model with
\beta_i=\gamma\z_i but no random error on this)
For ease of interpretation, I would let one or two
coefficients vary with z but keep other constant
over all the countries - put in fixed effects for
country here
d) now repeat c with lmer (remembering
link="logit") and figuring out how to specify
equations to do what you need - do for fixed and
random effects
What do you get
look at the predicitons of the random terms
II redo I but with the radon data - do for one
state other than what Gelman did
III Do an analysis on all the levels of the radon data