** read in the data **
use "C:\Documents and Settings\ejohnson\My Documents\MultilevelModels\cognitive.dta", clear

** exploratory **

** how many schools, how many treatments, and how many individuals within each school and each treatment group **
** need the wide format data to count individuals **
reshape wide arithmetic vmeaning dstotal ravens, i(id) j(rn)

tab schoolid
tab trt

table schoolid trt

** reshape back to long format for the analysis **
reshape long arithmetic vmeaning dstotal ravens, i(id) j(rn)

sort trt
by trt: summ  ravens arithmetic vmeaning dstotal

** OLS regression **
xi:reg ravens i.trt

** 3-level random intercept model **
xi:gllamm ravens i.trt, i(id school) adapt

xi:gllamm ravens i.trt age_at_time0 gender ses1 head_circ readtest writetest, i(id school) adapt

** 2-level random intercept model **
xi:gllamm ravens i.trt age_at_time0 gender ses1 head_circ readtest writetest, i(id) adapt

** individual longitudinal change **
sort id rn
graph twoway (line ravens rn, connect(ascending)), by(trt) xtitle("Time (weeks)") ytitle("Raven's score")

** random effect model **
xi:gllamm ravens i.trt rn age_at_time0 gender ses1 head_circ readtest writetest, i(id)

gen cons = 1
eq int1: cons
eq rnslope: rn
xi:gllamm ravens i.trt rn age_at_time0 gender ses1 head_circ readtest writetest, i(id) nrf(2) eqs(int rnslope) adapt