version 9.2

log using C:\Teaching\IGP\log\11.multiple.log, replace

* Chapter 11 Multiple regression
* modified on Feb. 04, 2007 by Leena Choi

use "C:\Teaching\IGP\data\EMS\perulung.dta", clear

***************************************************************************************
* Section 11.2 Multiple regression with two exposure variables
***************************************************************************************
* regress fev1 on age and height
regress fev1 age height

* calculate F-value using the MS
display 12.8193006/.052479971

* calculate F-value using the SS
display (25.6386013/2)/(33.2198218/633)

* calculate r-squared using the SS : r-squared = SSmodel/SStotal
display 25.6386013/58.858423

***************************************************************************************
* Section 11.3 Multiple regression with categorical exposure variables
***************************************************************************************
* regress fev1 on male
regress fev1 male

* put all label and define values
label variable male "Gender"
label define sexlab 0 "Girls" 1 "Boys"
label values male sexlab

* compare with ttest
ttest fev1, by(male)

* regress fev1 on age and male
regress fev1 age male

* regress fev1 on age, height and male
regress fev1 age height male

***************************************************************************************
* Section 11.6 Relationship between multiple regression and ANOVA
***************************************************************************************
use "C:\Teaching\IGP\data\hemoANOVA.dta", clear

* put informative label
label variable hemo "Hemoglobin (g/dL)"
label variable type "Type of sickle cell disease"
label define typelab 1 "Hb SS" 2 "Hb S/beta" 3 "Hb SC"
label values type typelab

* generate 2 dummy variable
generate type2 = 0
generate type3 = 0
replace type2 = 1 if type==2
replace type3 = 1 if type==3

regress hemo type2 type3

* compare with the result from one-way anova
oneway hemo type

***************************************************************************************

log close