age: Age

Description

age is the age of the person record.

Availability

Sample	Years
Basic	All years
May	All years
ORG	All years

EPI top codes the maximum age to 80 for consistency in maximum ages across years. Additionally, in the basic monthly files, the minimum age is 14 from 1976 to 1981 and 0 from 1982 to the present.

Values

Value	Label
80	80+

Detailed comments

The minimum age of the sample in the CPS extracts is consistent since 1982. However, in 1985, about 16-17% of observations ages 0-13 have zero values for finalwgt. The zero weight values in that year explain the spike in (sample-weighted) age shares illustrated above.

Code

Variable creation

********************************************************************************
* Age
********************************************************************************

capture rename age orig_age
gen int age = .

if $marchcps == 1 {
    if tm(1962m1) <= $date & $date <= tm(1997m12) {
        replace age = orig_age
    }

    if tm(1998m1) <= $date {
        replace age = a_age
    }
}

if $monthlycps == 1 | $maycps == 1 {
    if tm(1973m1) <= $date & $date <= tm(1993m12) {
        replace age = orig_age
    }
    if tm(1994m1) <= $date & $date <= tm(2012m4) {
        replace age = peage
    }
    if tm(2012m5) <= $date {
        replace age = prtage
    }
}

replace age = . if age < 0
recode age (80/max = 80)

cap lab drop age
lab def age 80 "80+"
lab val age age
lab var age "Age"
notes age: 1976-1993 Unicon: age
notes age: 1994-2012 CPS: peage
notes age: 2013-present CPS: prtage
notes age: top-coded at 80 for consistency across years
notes age: 1976-1981 range 14-80
notes age: 1982-present range 0-80

Figure creation

* restrict to sample when ages 0+ are available
keep if year >= 1982

gen byte child = age >= 0 & age <= 15
gen byte young = age >= 16 & age <= 24
gen byte early = age >= 25 & age <= 39
gen byte mid = age >= 40 & age <= 54
gen byte late = age >= 55 & age <= 64
gen byte old = age >= 65 & age <= 80

egen byte total = rowtotal(child young early mid late old)
assert total == 1

gcollapse (mean) child young early mid late old [pw=finalwgt], by(year) fast
foreach var of varlist child young early mid late old {
    replace `var' = `var' * 100
    sum `var' if year == 1987
    local `var'yvalue = r(mean)
    local `var'xvalue = 1987
}
local childyvalue = `childyvalue' - 1.1
local youngyvalue = `youngyvalue' + 0.9
local earlyyvalue = `earlyyvalue' + 1.0
local midyvalue = `midyvalue' + 3.0
local lateyvalue = `lateyvalue' - 1.9
local oldyvalue = `oldyvalue' - 0.7

local color1 228 26 28
local color2 55 126 184
local color3 77 175 74
local color4 152 78 163
local color5 255 127 0
local color6 166 86 40


line child young early mid late old year, ///
legend(off) ///
xlabel(1985(5)2020) ///
ylabel(5(5)20 25 "25%", angle(0) gmin gmax) ///
yscale(r(5 27)) ///
xtitle("") ytitle("") ///
lcolor("`color1'" "`color2'" "`color3'" "`color4'" "`color5'" "`color6'") ///
graphregion(color(white)) plotregion(color(white)) ///
title("Share of civilian population by age group", size(medium)) ///
text(`childyvalue' `childxvalue' "Ages 0-15", color("`color1'") placement(r)) ///
text(`youngyvalue' `youngxvalue' "Ages 16-24", color("`color2'") placement(r)) ///
text(`earlyyvalue' `earlyxvalue' "Ages 25-39", color("`color3'") placement(r)) ///
text(`midyvalue' `midxvalue' "Ages 40-54", color("`color4'") placement(r)) ///
text(`lateyvalue' `latexvalue' "Ages 55-64", color("`color5'") placement(r)) ///
text(`oldyvalue' `oldxvalue' "Ages 65+", color("`color6'") placement(r))
graph export ${variableimages}age_titleimage.svg, replace

/***
The minimum age of the sample in the CPS extracts is consistent since 1982. However, in 1985, about 16-17% of observations ages 0-13 have zero values for finalwgt. The zero weight values in that year explain the spike in (sample-weighted) age shares illustrated above.
***/