**************************************************************************
* exer5-1-do.txt                                                         *
*                                                                        *
* PURPOSE:  analysis of esphogeal cancer data                            *
*                                                                        *
* DATE:  01/05/11                                                        *
*                                                                        *
**************************************************************************
 
infile case age agegp tobgp tob logtob beer wine cider aperitif digest alc logalc  TuynsC-dat.txt

label variable case "case/control status"
label variable age "age"
label variable agegp "age group"
label variable tobgp "tobacco group"
label variable logtob "log(tob+1)"
label variable alc "total alcohol"
label variable logalc "log(alc+1)"

label define ylab 0 "control" 1 "case"
label values case ylab

label define alab 1 "25-34" 2 "35-44" 3 "45-54" 4 "55-64" 5 "65-74" 6 "75+"
label values agegp alab

drop if tob==99

xi:  logistic case i.agegp alc tob
logit
lrtest, saving(1)

***
*** do we need quadratic fnxs for either alc or tob?
***

generate alc2 = alc*alc
generate tob2 = tob*tob

xi:  logistic case i.agegp alc tob alc2
logit
lrtest, saving(2)

xi:  logistic case i.agegp alc tob tob2
logit
lrtest, saving(3)

lrtest, using(3) model(1)
lrtest, using(2) model(1)

***
*** descriptive statistics for the alcohol components
***

sort case

by case:  summarize beer
by case:  summarize wine
by case:  summarize cider
by case:  summarize aperitif
by case:  summarize digest

centile beer if case==0,     centile( 10 25 50 75 90 ) 
centile beer if case==1,     centile( 10 25 50 75 90 ) 

centile wine if case==0,     centile( 10 25 50 75 90 ) 
centile wine if case==1,     centile( 10 25 50 75 90 ) 

centile cider if case==0,    centile( 10 25 50 75 90 ) 
centile cider if case==1,    centile( 10 25 50 75 90 ) 

centile aperitif if case==0, centile( 10 25 50 75 90 ) 
centile aperitif if case==1, centile( 10 25 50 75 90 ) 

centile digest if case==0,   centile( 10 25 50 75 90 ) 
centile digest if case==1,   centile( 10 25 50 75 90 ) 

***
*** alcohol separated into components
***

xi: logistic case i.agegp tob beer wine cider aperitif digest
logit
lrtest, saving(4)

lrtest, using(4) model(1)


***
*** a look at the impact of age adjustment...
***

logistic case alc tob
logit