Document Sample
stata Powered By Docstoc
					    Introduction to Stata
       Paul von Hippel, Ph.D.
      Department of Sociology
       Ohio State University

with thanks to Stacy Woodard, Ph.D.
       Center for Biostatistics
        Ohio State University
                  Where you are
• Sociology Instruction Lab (SIL), 70 Derby Hall
  – used for computer courses
• not the Sociology Research Lab (SRL), in Bricker
  – used for computer research
        Logging in to the SIL
• Turn on computer
• Enter your user name
  – If your OSU email is
  – then your SIL user name is buendia.100
     • not the same as in SRL
• Enter your password
  – your first initial +
    last 4 digits of your SS# +
    your last initial
  – e.g., Jose Arcadio Buendia, SS# 100-10-0100
  – password j0100b
     • again not the same as in SRL
     Opening Stata (in the SIL)
• Lower left corner, click Start
• Choose
  – All programs
  – Analysis
  – Stata 9
            Logging commands
• Save your commands to a file
• You may want to re-use them later!

   – (Press return)
• called a ―command log‖
• or ―Do-file‖
        Opening Stata data files
• Stata has its own format for data files
   – extension *.dta
• Choose FileOpen
   – Go to S:\stata_intro\nmihs100.dta
      • S: in the SIL is not the same as S: in the SRL
• 100 records from
   – National Maternal and Infant Health Survey
   – in Stata format
Review window:                          Results window:
Past commands.                          Output and past commands
Click to paste in Command window

  Variables window:
                                        Command window:
  List of variables in open data set.
                                        Current command
  Click to paste in Command window.
                 Importing data
• Stata can also read tab-delimited ASCII text files
• Most other software (e.g., Excel)
  can write tab-delimited ASCII text files
• Let’s get data from Excel….
   – From Windows Start button
      • choose All ProgramsOffice Productivity
        Microsoft OfficeExcel
   – In Excel
      • choose FileOpen
          – find S:\stata_intro\example1.xls
      • choose FileSave As…Save as type: Text (tab delimited)
          – Save as X:\example1.txt
   – In Stata
      • In command window, type ―clear‖
        —gets current data out of memory
      • Choose FileImportASCII data created by a spreadsheet
          – Find X:\example1.txt
                    Examining data

                             selected   Hide   Delete
   Undo changes Sort by       column selected selected
                            to the end column columns         Close
  since last ―save‖selected                                   editor
―Save‖                                        or rows
• Move
  – smoke
  – and bwt
• to the first two columns
• Save your changes
• Close the editor
                     Saving data
• ―Preserve‖ saves only a temporary copy of the
  data file.
• The original data file is unaffected.
• To save a permanent data file,
   – Choose FileSave As…
   – Navigate to your X: drive
      • X: is where you should save things
      • X: in the SIL is not the same as X: in the SRL
   – Save as ―my_example1.dta‖
               Labeling variables
• To add a descriptive label to a variable
  – DataLabelsLabel variable
• Add these labels to these variables:
  – bwt : ―Birth weight, in grams‖
  – smoke : ―Did mother smoke during pregnancy?‖
                           Labeling values
• Many variables are dummy variables
   – two values: 0 and 1
       • e.g., ―Did the mother smoke?‖ Yes (1) or no (0).
• To add labels to dummy values
   – DataLabelsLabel ValuesDefine or Modify Value Labels
   – Define label name: ―dummy‖
   – Add values
       • 1 means ―yes‖
       • 0 means ―no‖
• Now tell Stata that smoke is a dummy variable
   – DataLabelsLabel ValuesAssign value label to variable
• Look at smoke in the Data Editor
   – and double-click it
           Creating a new variable
• According to the National Institutes of Health,
  – low birth weight (LBW)
     • < 2500 grams (5.5 pounds)
• Let’s create a dummy variable for LBW
• Data
  Create or change variable
  Create a new variable
• Create dummy variables for
  – very low birth weight (VLBW)
     • <2000 grams (4.4 pounds)
  – extremely low birth weight (ELBW)
     • <1500 grams (3.3 pounds)
• Look at the new variables in the Data Editor
• Give the new variables appropriate labels
  – e.g., LBW: ―Did the baby have low birth weight?‖
• NB. Stata is case-sensitive
  – ELBW is not the same as elbw
               Analyzing data
• Are smokers more likely to have LBW babies?
• Statistics
  Summaries, tables, and tests
  Two-way tables
             Analyzing data
• Are smokers’ babies lighter, on average?
• StatisticsSummaries…
  Summary statistics
  Summary statistics
             Formatting output
• In Stata
  – Highlight results
  – EditCopy table
• Open Excel
  – EditPaste
  – Tidy things up
  – Later move to Word processor
              Leaving Stata
• FileExit
               Resuming Stata
• Launch Stata:
• Old commands are in the do-file. To see it,

   – In do-file editor
      • FileOpen
      • Find X:\
                   Running a Do-file
• To run your old commands,
  – FileDo…
  – Find ―X:\‖
• It probably won’t run all the way through.
• Some commands don’t make sense any more
  – e.g., creating a file or variable that already exists
• Other commands may pause and wait for action from you
  – e.g., ―edit‖ waits for you to close the Data Editor
• Edit the do-file until it runs all the way
• This will also familiarize you with the typed
  versions of some commands.
                  To learn more
• Help menu (or help command)
   – e.g., in Command window type ―help cmdlog‖
• Tutorial command
   – tutorial ttest
   – tutorial graphics
• Manuals (in SRL)
   –   Getting Started
   –   User’s Guide
   –   Reference Manuals (several volumes)
   –   Stata Graphics Manual

Shared By: