Short Guides to Microeconometrics Prof. Dr.

Kurt Schmidheiny
Fall 2014 Universität Basel

A Short Guide to Stata 12

1 Introduction 2

2 The Stata Environment 2

3 Where to get help 3

4 Additions to Stata 3

5 Opening and Saving Data 4

6 Importing Data from Excel 5

7 Data Manipulation 6

8 Descriptive Statistics 9

9 Graphs 10

10 OLS Regression 11

11 Log Files 12

12 Do-Files 13

13 Important Functions and Operators 15

Version: 17-9-2014, 22:19

A Short Guide to Stata 12 2

1 Introduction

This guide introduces the basic commands of Stata. More commands are
described in the respective handouts.
All commands are shown using specific examples. Stata commands are
set in Courier; example specific datafiles , variables , etc. are set in
italics while built-in Stata functions and operators are upright.

2 The Stata Environment

When you start Stata, you will see the following windows: the Command
window where you type in your Stata commands, the Results window
where Stata results are displayed, the Review window where past Stata
commands are displayed and the Variables window which list all the vari-
ables in the active datafile.
The active datafile can be browsed (read-only) in the Browser window,
which is activated from the menu Data/Data browser or by the command

browse

The Editor window allows to edit data either by directly typing into the
editor window or by copying and pasting from spreadsheet software:

edit

Since version 8, Stata has implemented every command (except the pro-
gramming commands) as a dialog that can be accessed from the menus.
This makes commands you are using for the first time easier to learn as
the proper syntax for the operation is displayed in the Review window.
3 Short Guides to Microeconometrics

3 Where to get help

The printed Stata User’s Guide is an introduction into the capabilities

and basic concepts of Stata. The printed Stata Base Reference Manual
provides systematic information about all Stata commands. It is also
often an excellent treatise of the implemented statistical methods.
The online help in Stata describes all Stata commands with their options.
You can start the online with the help command. For example,

help correlate

describes the use of the command to calculate the correlation between

two or more variables. The online does not itself explain the statistical
methods. However, there is as link indicated by [R] in the top section
which opens the pdf version of the corresponding Reference manual.
If you don’t know the exact expression for the command, you can search
the Stata documentation by, for example,

search covariance

4 Additions to Stata

Many researchers provide their own Stata programs on Stata’s webpage.

You can search these resources from within Stata. For example,

net search unitroot

lists implementations of unit roots tests described in the Stata Journal

(SJ), the old Stata Technical Bulletin (STB) or submitted by users.
A Short Guide to Stata 12 4

5 Opening and Saving Data

Open an existing Stata datafile, for example mydata.dta, by

use mydata, clear

where the option clear removes a previously opened data set from the
Stata memory.
Stata provides a long series of example datasets at http://www.stata-press.
com/data/r12/. These dataset can directly be opened by, for example,

webuse lifeexp, clear

An open datafile is saved by, for example,

save mynewdata

in the file mynewdata.dta. If no filename is specified, the name under

which the data was last known is used.
Stata will look for data or save data in the current directory. Stata reports
the name and contents of the current directory by the command

dir

The current directory can be set to a particular drive and directory. For
example by

cd "/Users/kurt/Documents "
cd "C:\Users\kurt\Documents "

on Mac OS or Windows, respectively.

You may encounter memory problems when you open large datafiles. See
help memory on how to proceed in this case.
5 Short Guides to Microeconometrics

6 Importing Data from Excel

There are many ways to import data into Stata. Since version 12, data
can be conveniently imported from Excel with the menu File/Import/Excel
spreadsheet or the command import excel.
The following section shows a reliable way that can also be used in older
versions.
Prepare the data in Excel for conversion:

• Make sure that missing data values are coded as empty cells or as
numeric values (e.g., 999 or -1). Do not use character values (e.g -,
N/A) to represent missing data.

• Make sure that there are no commas in the numbers. You can
change this under Format menu, then select Cells... .

• Make sure that variable names are included only in the first row of
your spreadsheet. Variable names should be 32 characters or less,
start with a letter and contain no special characters except ‘ ’.

Under the File menu, select Save As... . Then Save as type Text(tab
delimited). The file will be saved with a .txt extension.
Start Stata, then issue the following command:

insheet using mydata.txt, clear

where mydata.txt is the name of the tab-delimited file. The option clear
removes a previously opened data set from the memory.
A Short Guide to Stata 12 6

7 Data Manipulation

New variables are created by the generate command. Existing variables

can be changed by the replace command.
For example:

webuse lifeexp
generate gnppc2 = gnppc ^ 2

generates a new variable gnppc2 with the square of the gross national
product (gnp).

generate lpopgrowth = log(popgrowth )

generates a new variable lpopgrowth with the natural log of population

growth. The system missing value code ‘.’ is assigned for observations
where the natural log is not defined.

generate rich = 1 if gnppc >= 20000

replace rich = 0 if gnppc < 20000

generates a new dummy variable rich taking the value one if the gnp is
greater or equal than 20000, zero if it is below and missing if it is unknown.

generate rich = gnppc >= 20000 if !missing(gnppc )

does the same. Note Stata returns true (1) for the conditional statement
gnppc >= 20000 if gnppc is missing. This is a very unfortunate feature
of Stata and the source of many errors.
The command egen extends the functionality of generate. For example

egen mgnppc = mean(gnppc )

7 Short Guides to Microeconometrics

creates a new variable mgnppc containing the (constant) mean of gnp for
all observations (countries). See the section 13 for more functions in egen.
Both the generate and the egen command allow the by varlist prefix
which repeats the command for each group of observations for which the
values of the variables in varlist are the same. For example,

sort region
by region : egen reggnppc = mean(gnppc )

generates the new variable reggnppc containing for each country the mean
of gnp across all countries within the same world region.
The recode command is a convenient way to exchange the values of or-
dinal variables. For example,

generate america = region

recode america (1=0) (2=1) (3=1)

will produce a new variable america where countries in South America

(region ==3) and North America (region ==2) are assigned one and
countries in Europe and Asia (region ==1) are assigned zero.
The following system variables (note the ‘ ’) may be useful:
n contains the number of the current observation.
N contains the total number of observations in the dataset.
pi contains the value of pi to machine precision.
You can delete variables from the dataset by either specifying the variables
to be dropped or to be kept. For example,

drop gnppc2 mgnppc

drops the two variables gnppc2 and mgnppc ;

A Short Guide to Stata 12 8

keep region country popgrowth lexp gnppc

drops all but the listed variables.

You can delete observations from the dataset by specifying the observa-
tions to be dropped (or kept) in a logical expression or by specifying the
first and last observation. For example,

drop if region == 3

drops all observations from South America;

keep in 6/20

keeps only the 15 observations from row 6 to row 20.

You can arrange the observations of the current dataset in ascending order
with respect to one or more specific variables. For example,

sort region lexp

sorts the countries within regions by life expectancy. The order of vari-
ables in the current dataset is changed with, for example,

order country region popgrowth

moves the variable country to the front.

You can conveniently reduce a datasets to aggregate statistics such as the
mean (see help collapse for more). For example,

collapse (mean) lexp gnppc (sd) sdgnppc=gnppc, by(region )

produces a dataset with 3 observations, one for each world region. The
new dataset contains the mean of the variables lexp and gnppc as well
as the standard deviation of gnppc .
9 Short Guides to Microeconometrics

8 Descriptive Statistics

A description of the variables in the dataset is produced by the commands

describe and codebook.
Further commands report selected statistics. For example,

webuse lifeexp
summarize lexp gnppc

reports univariate summary statistics of the variables lexp gnppc .

tabulate safewater, missing

reports the frequency counts for safewater. The missing option requests
that the number of missing values is also reported.

correlate popgrowth lexp gnppc

correlate popgrowth lexp gnppc, covariance

displays the correlation and, respectively, the covariance matrix between

popgrowth , lexp and gnppc .

generate rich = gnppc >= 20000 if !missing(gnppc )

tabulate rich region, col chi2

produces a two-way table of absolute and relative frequencies counts along

with Pearson’s chi-square statistic.

ttest lexp if region == 2 | region == 3, by(region ) unequal

performs a two-sample t-test of the null hypothesis that the life expectancy
in North and South America is identical. The option unequal indicates
that the two groups are not assumed to have equal variances.
A Short Guide to Stata 12 10

9 Graphs

Stata has many commands to draw figures. For example,

webuse lifeexp
scatter lexp gnppc

draws a scatter plot of the variable lexp (y-axis) against gnppc (x-axis).

line lexp gnppc, sort

draws a line graph, i.e. scatter with connected points.

scatter lexp gnppc || lfit lexp gnppc

draws a scatter plot with regression line. A histogram with relative fre-
quencies is called with, for example,

histogram gnppc
11 Short Guides to Microeconometrics

10 OLS Regression

The multiple linear regression model is estimated by OLS with the regress
command For example,

webuse auto
regress mpg weight displacement

regresses the mileage(mpg ) of a car on weight and displacement . A

constant is automatically added if not suppressed by the option noconst

regress mpg weight displacement, noconst

Estimation based on a subsample is performed as

regress mpg weight displacement if weight > 3000

where only cars heavier than 3000 lb are considered. The Eicker-Huber-
White covariance is reported with the option robust

regress mpg weight displacement, vce(robust)

F -tests for one or more restrictions are calculated with the post-estimation
command test. For example

test weight displacement

tests H0 : β1 = 0 and β2 = 0 against HA : β1 6= 0 or β2 6= 0.

New variables with residuals and fitted values are generated by

predict uhat if e(sample), resid

predict pricehat if e(sample)
A Short Guide to Stata 12 12

11 Log Files

A log file keeps a record of the commands you have issued and their results
during your Stata session. You can create a log file with, for example

log using mylog.txt, replace text

where mylog.txt is the name of the resulting log file. The append option
adds more information to an existing file, whereas the replace option
erases anything that was already in the file. Full logs are recorded in one
of two formats: SMCL (Stata Markup and Control Language) or text
(ASCII). The default is SMCL, but the option text changes that.
A command log contains only your commands, for example

cmdlog using mycmd.txt

Both type of log files can be viewed in the Viewer:

view mylog.txt

You can temporarily suspend, resume or stop the logging with the com-
mands:

log on
log off
log close
13 Short Guides to Microeconometrics

12 Do-Files

A “do”-file is a set of commands just as you would type them in one-by-

one during a regular Stata session. Any command you use in Stata can be
part of a do file. The default extension of do-files is .do, which explains
its name. Do-files allow you to run a long series of commands several
times with minor or no changes. Furthermore, do-files keep a record of
the commands you used to produce your results.
To edit a do file, just click on the icon (like an envelope) in the toolbar.
To run this file, save it in the do-file editor, for example as mydofile.do
and issue the command:

do mydofile.do

You can also click on the Do current file icon in the do-file editor to run
the do file you are currently editing.
Comments are indicated by a * at the beginning of a line. Alternatively,
what appears inside /* */ is ignored. The /* and */ comment delimiter
has the advantage that it may be used in the middle of a line.

* this is a comment
generate x = 2*y /* this is another comment*/ + 5

Hitting the return key tells Stata to execute the command. In a do file,
the return key is at the end of every line, and restricts commands to be on
the same line with a maximum of 255 characters. In many cases, (long)
commands are more clearly arranged on multiple lines. You can tell Stata
that the command is longer than one line by using the

#delimit ;

command in the beginning of your do-file. The following Stata commands

A Short Guide to Stata 12 14

are now terminated by a ‘;’. An example do-file:

capture log using mincer, replace

#delimit ;
use schooling.dta, clear ;
* generate a proxy for experience ;
generate exp = age - educ - 6 ;
* estimate the Mincer equation ;
regress
lnwage educ exp exp2 female
/* change the significance level to 0.01 */
, level(99) ;
log close ;

⇒ Note that lines with comments also need to be terminated by ‘;’. Oth-
erwise the following command will not be executed.
15 Short Guides to Microeconometrics

13 Important Functions and Operators

Some Mathematical Expressions

abs(x) returns the absolute value of x.

exp(x) returns the exponential function of x.
int(x) returns the integer by truncating x towards zero.
ln(x), log(x) returns the natural logarithm of x if x>0.
log10(x) returns the log base 10 of x if x>0.
max(x1,...,xn) returns the maximum of x1, ..., xn.
min(x1,...,xn) returns the minimum of x1, ..., xn.
round(x) returns x rounded to the nearest whole number.
round(x,y) returns x rounded to units of y.
sign(x) returns -1 if x<0, 0 if x==0, 1 if x>0.
sqrt(x) returns the square root of x if x>=0.

Logical and Relational Operators

& and | or
! not ∼ not
> greater than < less than
>= greater or equal <= smaller or equal
== equal != not equal

Some Probability distributions and density functions

normal(z) cumulative standard normal distribution

normalden(z) returns the standard normal density
normalden(z,m,s) normal density with mean m and stand. deviation s
invnormal(p) inverse cumulative standard normal distribution

Similar commands are available for a variety of distribution functions.

A Short Guide to Stata 12 16

Some Functions in egen

fill(numlist )
creates a variable of ascending or descending numbers or complex
repeating patterns. See help numlist for the numlist notation.

max(varname ) (allows by varlist :)

creates a constant containing the maximum value of varname.

mean(varname )
creates a constant containing the mean of varname.

median(varname ) (allows by varlist :)

creates a constant containing the median of varname.

min(varname ) (allows by varlist :)

creates a constant containing the minimum value of varname.

rowmax(varlist )
gives the maximum value in varlist for each observation (row). Equals
max(var1, var2, ... ) in the generate command.

rowmean(varlist )
creates the (row) means of the variables in varlist for each observation
(row).

rowmin(varlist )
gives the minimum value in varlist for each observation (row).

sd(varname ) (allows by varlist :)

creates a constant containing the standard deviation of varname.

sum(varname ) (allows by varlist :)

creates a constant containing the sum of varname.