Introduction to Stata

December 26, 2016

Have you heard about Stata?

What is Stata?

Statistical software
Has a user interface and code
Most statistical methods are implemented
It's licensed software, so it's not free

Why Stata?

User interface: means users do not need to know how to write code
It is great for mixed models, survey data analysis, etc
Can be used for reproducible work via do files
Researchers can contribute their modules: there's even a Stata journal

Main components

Data browser
- View your data
- Can edit but not recommended

Help files: very detailed and interconnected
Can import data from many file types
Intuitive menu
Console: shows the latest code
Log: shows the latest results
do file editor: make your work reproducible

Are you ready to start using Stata?

First, type the following command (or copy paste it)

sysuse autos

What do you get?

sysuse autos


. sysuse autos
file "autos.dta" not found
r(601);

end of do-file
r(601);

So, what went wrong?

Lets see if the error message gives us a hint as to what could have gone wrong.

The filename you specified cannot be found.  Perhaps you
mistyped the name, or it may be on another CDor directory.

Ohh, it could be that we incorrectly typed the name! Try this next:

sysuse auto

Now lets explore the data

Command review

Basic syntax

What do you want to do?

Described by an action
Think of it as a verb
For example, load a data set included in Stata with sysuse

What are you going to use for that action?

That's the actual data that you will use
Think of it as the subject
For example, the auto data set

Keep track of your work

Do you remember all the options in a user interface you used 6 months ago?
It's important to keep track of your work!
Useful when learning so you can revisit what you did: see what worked, what didn't
Two main options in Stata: log files and do files

Log files

These files keep track of everything:

The commands you used
The commands that specify what you did with the user interface
The output you generated: tables, results
They do not save images

Start a log for our session

Example log file

Do files

These files contain commands only
They are an executive summary of what you did
Great for analyses you want to share
You can include comments describing the logic of what you are doing
You can execute them to run commands
Cleaner than log files.
Open the do file and paste the command to load the auto data set.

Example do file

Exercise

Use the Afghanistan data from the Worldbank and make a plot of the agricultural land (in squared kilometers).

Start by loading the data!

No cheating!!!

Don't look at the next slides! Try first!

What variable is it?

It's AGLNDAGRIK2

use afg_worldbank_2016.dta, clear
graph box AGLNDAGRIK2

Looks pretty easy, right? Just 2 lines of code.

The original data was a bit more messy! Transforming it to something we can use is called data cleaning or data tidying.

Stata syntax: options

Many Stata commands have options that give us more fine tuned control over what we want to do
You can find them also via the user interface: many are checkboxes
Think of options as adjectives
Examples: title of a graph, which data to use, colors
They normally come after a comma
You can find the options in the Stata help

Options example

use "HH Listing.dta"
graph box q105 if q103 == 1, by(q104)

graph box q105 if q103 == 1,ytitle(Age of Head of Household)/*
    */ by(, title(Age of Head of Household by Gender) /*
    */ subtitle(Afghanistan)) by(q104)

Stata help

It comes in two shapes: help pages for commands and PDF documents
Help pages are more direct
Sometimes it's easier to scroll down to the examples section
The help pages are interconnected, so use the hyperlinks
Check out the help page for graph box: help graph box

Exercise continued

Improve the plot you made earlier as if you were to show the distribution of the agricultural land of Afghanistan in the media (news) or a journal.

graph hbox AGLNDAGRIK2, title(Afghanistan agricultural land) /*
    */ subtitle(1961 to 2013) note(Worldbank 2016) scheme(sj)

See all the entries for a variable: list

list AGLNDAGRIK2


. use afg_worldbank_2016.dta, clear
(Written by R.              )

. list AGLNDAGRIK2

     +----------+
     | AGLNDA~2 |
     |----------|
  1. |        . |
  2. |   377000 |
  3. |   377600 |
  4. |   378100 |
  5. |   378730 |
     |----------|
  6. |   378750 |
  7. |   379130 |
  8. |   379790 |
  9. |   379800 |
 10. |   379960 |
     |----------|
 11. |   380060 |
 12. |   380360 |
 13. |   380460 |
 14. |   380480 |
 15. |   380480 |
     |----------|
 16. |   380480 |
 17. |   380480 |
 18. |   380500 |
 19. |   380500 |
 20. |   380490 |
     |----------|
 21. |   380490 |
 22. |   380530 |
 23. |   380540 |
 24. |   380540 |
 25. |   380540 |
     |----------|
 26. |   380540 |
 27. |   380540 |
 28. |   380450 |
 29. |   380400 |
 30. |   380400 |
     |----------|
 31. |   380400 |
 32. |   380300 |
 33. |   380300 |
 34. |   379340 |
 35. |   378130 |
     |----------|
 36. |   377530 |
 37. |   377520 |
 38. |   377900 |
 39. |   378670 |
 40. |   377530 |
     |----------|
 41. |   377530 |
 42. |   377530 |
 43. |   377530 |
 44. |   379100 |
 45. |   379110 |
     |----------|
 46. |   379100 |
 47. |   379100 |
 48. |   379100 |
 49. |   379100 |
 50. |   379100 |
     |----------|
 51. |   379100 |
 52. |   379100 |
 53. |   379100 |
 54. |   379100 |
 55. |        . |
     |----------|
 56. |        . |
 57. |        . |
     +----------+

codebook: main information

codebook AGLNDAGRIK2


. use afg_worldbank_2016.dta, clear
(Written by R.              )

. codebook AGLNDAGRIK2

-----------------------------------------------------------------------------------------
AGLNDAGRIK2                                                    Agricultural land (sq. km)
-----------------------------------------------------------------------------------------

                  type:  numeric (double)

                 range:  [377000,380540]              units:  10
         unique values:  28                       missing .:  4/57

                  mean:    379405
              std. dev:   1110.28

           percentiles:        10%       25%       50%       75%       90%
                            377530    378750    379340    380480    380530

summarize: univariate statistical summary

summarize AGLNDAGRIK2


. use afg_worldbank_2016.dta, clear
(Written by R.              )

. summarize AGLNDAGRIK2

    Variable |        Obs        Mean    Std. Dev.       Min        Max
-------------+---------------------------------------------------------
 AGLNDAGRIK2 |         53    379404.5    1110.278     377000     380540

summarize AGLNDAGRIK2, detail


. use afg_worldbank_2016.dta, clear
(Written by R.              )

. summarize AGLNDAGRIK2, detail

                 Agricultural land (sq. km)
-------------------------------------------------------------
      Percentiles      Smallest
 1%       377000         377000
 5%       377530         377520
10%       377530         377530       Obs                  53
25%       378750         377530       Sum of Wgt.          53

50%       379340                      Mean           379404.5
                        Largest       Std. Dev.      1110.278
75%       380480         380540
90%       380530         380540       Variance        1232718
95%       380540         380540       Skewness      -.5865218
99%       380540         380540       Kurtosis       2.045356

describe: similar to `codebook`

describe AGLNDAGRIK2


. use afg_worldbank_2016.dta, clear
(Written by R.              )

. describe AGLNDAGRIK2

              storage   display    value
variable name   type    format     label      variable label
-----------------------------------------------------------------------------------------
AGLNDAGRIK2     double  %9.0g                 Agricultural land (sq. km)

tabulate: make a table

tabulate AGLNDAGRIK2


. use afg_worldbank_2016.dta, clear
(Written by R.              )

. tabulate AGLNDAGRIK2

Agricultura |
l land (sq. |
        km) |      Freq.     Percent        Cum.
------------+-----------------------------------
     377000 |          1        1.89        1.89
     377520 |          1        1.89        3.77
     377530 |          5        9.43       13.21
     377600 |          1        1.89       15.09
     377900 |          1        1.89       16.98
     378100 |          1        1.89       18.87
     378130 |          1        1.89       20.75
     378670 |          1        1.89       22.64
     378730 |          1        1.89       24.53
     378750 |          1        1.89       26.42
     379100 |         10       18.87       45.28
     379110 |          1        1.89       47.17
     379130 |          1        1.89       49.06
     379340 |          1        1.89       50.94
     379790 |          1        1.89       52.83
     379800 |          1        1.89       54.72
     379960 |          1        1.89       56.60
     380060 |          1        1.89       58.49
     380300 |          2        3.77       62.26
     380360 |          1        1.89       64.15
     380400 |          3        5.66       69.81
     380450 |          1        1.89       71.70
     380460 |          1        1.89       73.58
     380480 |          4        7.55       81.13
     380490 |          2        3.77       84.91
     380500 |          2        3.77       88.68
     380530 |          1        1.89       90.57
     380540 |          5        9.43      100.00
------------+-----------------------------------
      Total |         53      100.00

tabulate: include missing observations

tabulate AGLNDAGRIK2, missing


. use afg_worldbank_2016.dta, clear
(Written by R.              )

. tabulate AGLNDAGRIK2, missing

Agricultura |
l land (sq. |
        km) |      Freq.     Percent        Cum.
------------+-----------------------------------
     377000 |          1        1.75        1.75
     377520 |          1        1.75        3.51
     377530 |          5        8.77       12.28
     377600 |          1        1.75       14.04
     377900 |          1        1.75       15.79
     378100 |          1        1.75       17.54
     378130 |          1        1.75       19.30
     378670 |          1        1.75       21.05
     378730 |          1        1.75       22.81
     378750 |          1        1.75       24.56
     379100 |         10       17.54       42.11
     379110 |          1        1.75       43.86
     379130 |          1        1.75       45.61
     379340 |          1        1.75       47.37
     379790 |          1        1.75       49.12
     379800 |          1        1.75       50.88
     379960 |          1        1.75       52.63
     380060 |          1        1.75       54.39
     380300 |          2        3.51       57.89
     380360 |          1        1.75       59.65
     380400 |          3        5.26       64.91
     380450 |          1        1.75       66.67
     380460 |          1        1.75       68.42
     380480 |          4        7.02       75.44
     380490 |          2        3.51       78.95
     380500 |          2        3.51       82.46
     380530 |          1        1.75       84.21
     380540 |          5        8.77       92.98
          . |          4        7.02      100.00
------------+-----------------------------------
      Total |         57      100.00

count number of observations

count if AGLNDAGRIK2 == .


. use afg_worldbank_2016.dta, clear
(Written by R.              )

. count if AGLNDAGRIK2 == .
  4

generate: create a new variable

generate agrilog = log10(AGLNDAGRIK2)


. use afg_worldbank_2016.dta, clear
(Written by R.              )

. generate agrilog = log10(AGLNDAGRIK2)
(4 missing values generated)

. summarize agrilog

    Variable |        Obs        Mean    Std. Dev.       Min        Max
-------------+---------------------------------------------------------
     agrilog |         53    5.579101     .001272   5.576341     5.5804

drop: delete variables

drop agrilog


. use afg_worldbank_2016.dta, clear
(Written by R.              )

. generate agrilog = log10(AGLNDAGRIK2)
(4 missing values generated)

. drop agrilog

You can use keep if the list of variables you want to retain is shorter than the list of variables you want to delete.

Tougher exercise

Calculate a t-test for a difference in means for the agricultural land between the even and the odd years.

Tip: remember that in math the modulus 2 is 0 for even values and 1 for odd values.


. use afg_worldbank_2016.dta
(Written by R.              )

. generate modulus = mod(Year, 2)

. label define moduluslabel 0 "Even" 1 "Odd"

. label values modulus moduluslabel

. ttest AGLNDAGRIK2, by(modulus)

Two-sample t test with equal variances
------------------------------------------------------------------------------
   Group |     Obs        Mean    Std. Err.   Std. Dev.   [95% Conf. Interval]
---------+--------------------------------------------------------------------
    Even |      26    379436.2    214.8517    1095.533    378993.7    379878.6
     Odd |      27    379374.1    220.2221    1144.308    378921.4    379826.7
---------+--------------------------------------------------------------------
combined |      53    379404.5    152.5084    1110.278    379098.5    379710.6
---------+--------------------------------------------------------------------
    diff |            62.07977    307.9249               -556.1052    680.2647
------------------------------------------------------------------------------
    diff = mean(Even) - mean(Odd)                                 t =   0.2016
Ho: diff = 0                                     degrees of freedom =       51

    Ha: diff < 0                 Ha: diff != 0                 Ha: diff > 0
 Pr(T < t) = 0.5795         Pr(|T| > |t|) = 0.8410          Pr(T > t) = 0.4205

Some common complicated tasks

Reshaping data sets: wide to long and viceversa
Merging data sets: Stata ultimately wants you to get all the data you need in a single table
Looping over a list to do a task
Working with strings
Working with dates: there are many different ways to specify time

Finding help

Try using the user interface then check the code that results from your choices
Use the help command or search the help (top right)
Google: lots of goodies in older mailing list posts
Check the UCLA Stata website

More information

The Statistics with Stata by Lawrence Hamilton you already have!
A quick rundown of the main commands in Stata by the UNC Caroline Population Center
A recording of an introduction to Stata by Dr Marie Diener-West
SPSS to Stata table
UCLA Stata Starter Kit