* do-file for home assignment 2 of vhm 811, Fall semester 2009
* Henrik Stryhn

version 10
use "a2.dta", clear

* Q2
ir morts smoker person_years

* Q3
gen agediff4=(age~=4)
ir morts agediff4 person_years if age==1 | age==4
ir morts agediff4 person_years if age==2 | age==4
ir morts agediff4 person_years if age==3 | age==4
ir morts agediff4 person_years if age==5 | age==4
ir morts agediff4 person_years if age==6 | age==4
ir morts agediff4 person_years if age==7 | age==4
gen ageg1=(age>1)
ir morts ageg1 person_years
gen ageg2=(age>2)
ir morts ageg2 person_years
gen ageg3 =(age>3)
ir morts ageg3 person_years
gen ageg4=(age>4)
ir morts ageg4 person_years
gen ageg5=(age>5)
ir morts ageg5 person_years
gen ageg6=(age>6)
ir morts ageg6 person_years

* Q4
* M-H estimation
ir morts smoker person_years, by(age)
ir morts smoker person_years if age>1, by(age)
* association age-mortality
ir morts agediff4 person_years if (age==1 | age==4) & smoker==0
ir morts agediff4 person_years if (age==2 | age==4) & smoker==0
ir morts agediff4 person_years if (age==3 | age==4) & smoker==0
ir morts agediff4 person_years if (age==5 | age==4) & smoker==0
ir morts agediff4 person_years if (age==6 | age==4) & smoker==0
ir morts agediff4 person_years if (age==7 | age==4) & smoker==0
* association age-smoking
gen count=round(pct_nonsm/100*total)
replace count=round((1-pct_nonsm/100)*total) if smoker==1
tabulate smoker age [fw=count], chi2 column

* Q5: test for homogeneity already computed under Q4

* Q6
tabulate smoker age [fw=count], chi2 row