Advanced Stata
Programming
{
Andrew Hicks
CCPR Statistics and Methods Core
Automating your work
1.Macros
2.Saved Results
3.Loops
4.Egen commands
5.Writing programs
6.Ado files
Macros
Assigns a string of text or a number to an abbreviation.
( a box you put text in)
local name content
local x 1
display `x’
(italics indication something you provide)
Macros
Assigns a string of text or a number to an abbreviation.
local nobs = 198
display “The number of observations equals: `nobs’”
display `nobs’ – 90
Local macros can be used only within the do-file in which
the are defined.
• When the program ends, the macro disappear.
local x 1
Global macros persist until you delete it or exit Stata.
global x 1
Macros
Assigns a string of text or a number to an abbreviation.
( a box you put text in)
regress price headroom trunk weight length
`size’
local size “headroom trunk weight length”
regress price `size’
regress mpg `size’
regress gear_ratio `size’ foreign turn
Results saved by Stata
Stata provides the number so you don’t have to enter
by hand.
summarize price
return list
generate price_centered = price – 6165.257
generate price_centered2 = price – r(mean)
Loops
Center 5 variables from Auto dataset: price, mpg, weight, length, turn, displacement
Manually:
summarize price
generate price_centered = price – r(mean)
summarize mpg
generate mpg_centered = mpg – r(mean)
summarize weight
generate weight_centered = weight – r(mean)
summarize length
generate length_centered = weight – r(mean)
summarize turn
generate turn_centered = turn – r(mean)
summarize displacement
generate displacement = displacement – r(mean)
Foreach Loops
Center 6 variables from Auto dataset: price, mpg, weight, length, turn, displacement
Loop (foreach):
foreach var in price mpg weight length turn displacement {
summarize `var’
generate `var’_centered2 = `var’ – r(mean)
}
1. summarize price
generate price_centered2 = price – r(mean)
2. summarize mpg
generate mpg_centered2 = mpg – r(mean)
.
.
6. summarize displacement
generate displacement_centered2 = displacement – r(mean)
Foreach Loops
Loop (foreach):
foreach depvar in weight length turn displacement {
regress `depvar’ price mpg headroom trunk
}
foreach depvar of varlist weight-displacement {
regress `depvar’ price mpg headroom trunk
}
foreach item in quest1 qeust2 qeust3 quest4 {
replace `item’=. if `item’ == 99
}
foreach item of varlist quest* {
replace `item’=. if `item’ == 99
}
Foreach Loops
Create your own foreach loop:
generate
generate
generate
generate
generate
generate
generate
generate
generate
generate
generate
generate
taxinc1 = inc1 *
taxinc2 = inc2 *
taxinc3 = inc3 *
taxinc4 = inc4 *
taxinc5 = inc5 *
taxinc6 = inc6 *
taxinc7 = inc7 *
taxinc8 = inc8 *
taxinc9 = inc9 *
taxinc10 = inc10
taxinc11 = inc11
taxinc12 = inc12
.10
.10
.10
.10
foreach lname in list {
.10
commands referring to lname
.10
}
.10
.10
.10
* .10
* .10
* .10
Forvalues Loops
gen
gen
.
.
gen
gen
hadInc1990 = (inc1990>0) if inc1990<.
hadInc1991 = (inc1991>0) if inc1991<.
hadInc2009 = (inc2009>0) if inc2009<.
hadInc2010 = (inc2010>0) if inc2010<.
Forvalues Loops
Loop (forvalues):
forvalues year=1990/2010 {
gen hadInc`year’=(inc`year’>0) if inc`year’<.
}
Forvalues Loops
Loop (forvalues):
forvalues year=1990/2010 {
gen hadInc`year’=(inc`year’>0) if inc`year’<.
}
forvalues year=1990(2)2010 {
gen hadInc`year’=(inc`year’>0) if inc`year’<.
}
foreach year of numlist 1980 1983 1990{
gen hadInc`year’=(inc`year’>0) if inc`year’<.
}
forvalues race=1/3 {
svy, subpop(if race==`race’): regress income age i.education
}
levelsof race, local(races)
foreach race in races {
svy, subpop(if race==`race’): regress income age i.education
}
Forvalues Loops
Create your own forval loop:
generate
generate
generate
generate
generate
generate
generate
generate
generate
generate
generate
generate
taxinc1 = inc1 *
taxinc2 = inc2 *
taxinc3 = inc3 *
taxinc4 = inc4 *
taxinc5 = inc5 *
taxinc6 = inc6 *
taxinc7 = inc7 *
taxinc8 = inc8 *
taxinc9 = inc9 *
taxinc10 = inc10
taxinc11 = inc11
taxinc12 = inc12
.10
.10
.10
.10
forvalues lname = range {
.10
commands referring to lname
.10
}
.10
.10
.10
* .10
* .10
* .10
Nested Loops
forval i=1/3 {
forval j=1/3 {
display “`i’, `j’”
}
}
Nested Loops
forval year = 1990/2010 {
foreach month in Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec {
gen hadInc`month’`year’ = (inc`month’`year’>0) if inc`month’`year’<.
}
}
Loops with counter
local counter = 0
foreach var in price mpg weight length turn displacement {
local
= `counter’ + 1
local counter
++counter
display “`counter’ - `var’”
}
local counter = `counter’ + 1
local ++counter
While Loops
foreach and forvalues loops repeat a command a set number of times:
forval i=1/5 {
display `i’
}
while loops repeat until a condition is no longer true:
local i 1
while `i’<=5 {
display `i++’
}
Egen commands
Take the average:
generate mean_score = (read + write + math + science)/4
But if any single score is missing, the total score will be missing
Solution:
egen mean_score = rowmean(read write math science)
egen mean_score = rowmean(score1 score2 score3 score4)
egen mean_score = rowmean(score*)
= rowtotal(score*)
= rowmax(score*)
= rowmin(score*)
Egen commands
rowmean( ) averages across variables
mean( ) averages across observations
egen mean_read = mean(read)
egen schooltype_mean_read = mean(read), by(schtyp)
Writing programs
A simple program:
program define hello
di “Hello World”
end
A more complicated program:
capture program drop hello
program define hello
di “Hello `1’”
end
An even more complicated program:
capture program drop hello
program define hello
di “Hello `0’”
end
Writing programs
A program to center variables:
program define demean
foreach var of local 0 {
quietly: sum `var'
replace `var'=`var'-r(mean)
}
end
Writing programs
A program to center variables:
program define demean
foreach var of local 0 {
capture confirm numeric variable `var'
if _rc==0 {
sum `var',meanonly
replace `var'=`var'-r(mean)
}
else di "`var' is not a numeric variable and cannot be demeaned."
}
end
Ado files
How to make an ado file:
1. Write a program in a do-file
2. Save the do-file with .ado extension
demean.ado
3. Put the .ado file in your personal ado directory
sysdir
PERSONAL: c:\ado\personal\
4. Use your ado program just like any other Stata command