*****pgm: hypothesis_geometric.do****

use "C:\nhanes\data\analysis_data.dta", clear

***create variable codes*****
gen     age1=1 if ridageyr >=20 & ridageyr <30
replace age1=2 if ridageyr >=30 & ridageyr <40
replace age1=3 if ridageyr >=40 & ridageyr <50
replace age1=4 if ridageyr >=50 & ridageyr <60
replace age1=5 if ridageyr >=60 & ridageyr <70
replace age1=6 if ridageyr >=70 & ridageyr <.

gen     race=1 if ridreth1==3
replace race=2 if ridreth1==4
replace race=3 if ridreth1==1
replace race=4 if ridreth1==2 | ridreth1==5

****format variables******
label define age1fmt 1 "20-29"
label define age1fmt 2 "30-39", add
label define age1fmt 3 "40-49", add
label define age1fmt 4 "50-59", add
label define age1fmt 5 "60-69", add
label define age1fmt 6 "70+", add

label define racefmt 1 "NH White", 
label define racefmt 2 "NH Black", add
label define racfmt  3 "Mex American", add
label define racefmt 4 "Other race/ethn", add

label define sexfmt 1 "male"
label define sexfmt 2 "female", add

label values age1 age1fmt
label values riagendr sexfmt
label values race racefmt

log using "c:\nhanes\log\hypothesis_geometric.log", replace

****specify survey design variables****
svyset [w=wtsaf4yr], psu(sdmvpsu) strata(sdmvstra) vce(linearized)

gen lnlbxtr=ln(lbxtr)

svy:mean lnlbxtr, subpop(if ridageyr >=20 & ridageyr <.)
ereturn display, eform(geo_mean)

svy:mean lnlbxtr, subpop(if ridageyr >=20 & ridageyr <.) over(riagendr)
ereturn display, eform(geo_mean)

svy:mean lnlbxtr, subpop(if ridageyr >=20& ridageyr <.) over(riagendr age1)
ereturn display, eform(geo_mean)

log close