Two Page Stata

Andrew Grogan-Kaylor

August 05, 2017

An introduction to Stata in 2 pages. Commands that you actually type into Stata are represented in monospace font. x and y refer to variables in your data. The treatment here is intended to be extremely brief, in order to create a kind of “cheat sheet” that can be presented in 2 pages. More documentation on any command is available in the printed or PDF Stata manuals, or by typing help command.

The Stata interface makes it extremely easy to do rapid interactive data analysis. Hit PAGE-UP to recall the most recent command, which you can then quickly edit and resubmit. The general idea of most Stata commands is command variables, options. Often it is not necessary to use any options since the authors of Stata have done such a good job of thinking about the defaults.

Use the DO FILE EDITOR to save Stata commands that you want to use again, and to create an “audit trail” of your work so that your workflow is documented and replicable. log using filename, replace will save a log file of your results. log close closes the log file.

Get Acquainted With Your Data

codebook x y will produce a nicely formatted codebook of selected variables, which is especially useful if you have added variable labels with the label variable command. codebook is especially useful for seeing how numerical values are associated with value labels. codebook by itself will list every variable in your data and generate a lot of [probably too much] output.

lookfor allows you to find variables that contain a specified keyword. This is especially useful in large data sets with many variables. Often abbreviated keywords are the most helpful. e.g. to find a poverty variable, type lookfor pov.

describe tells you about the contents of a specific variable. E.g. describe x y. describe, short will tell you very basic things about your data, including the number of observations in the data set, and the size of your data file.

Process Your Data

Data with missing values, often represented as negative numbers (e.g. -99, -9, -8) need to be recoded so that the missing values are represented as a missing value character (“.”) that Stata knows to exclude from calculations.

recode x (oldvalue = newvalue), generate(z) will recode a variable into a new variable, often a good idea.

recode _all (-99/-1 = .) will recode all negative numbers from -99 to -1 to missing for all variables in your data. recode x (7/9 = .) changes 7 through 9 to be missing for x. Indeed, recode will change specific values in your data to anything you want, not just missing values.

It is often convenient to rename your variables so that the variables have more intuitively understandable names e.g. rename x depression.

You can create new variables out of old variables using generate newvar = expression e.g. generate newvar = oldvar1 + oldvar2.

It is sometimes useful to sort your data. sort x will sort your data by the values of x.

Descriptive Statistics

summarize gives you basic descriptive statistics for a variable, such as the mean (average). Especially useful for continuous variables. E.g. summarize x y or summarize x y, detail.

tabulate gives you a frequency distribution for your variable. Especially useful for categorical variables. e.g. tabulate x.

Bivariate Statistics

Tabulating two categorical variables together gives you a cross-tabulation of those variables, e.g tabulate x y, row col chi2

pwcorr x y, sig gives you the pairwise correlation of two continuous variables.

oneway continous_var categorical_var, tabulate gives you a oneway ANOVA of a continuous variable over a categorical factor.

Multivariate Statistics

regress y x regresses y on x.
regress y x z regresses y on x and z. (Other regression commands follow a very similar format: command y x z but are beyond the purview of this 2 page guide.) regress y x i.z regresses y on x and z, treating x as continuous and z as a set of categorical indicator variables.

Stata’s factor variables allow one to generate interactions and indicator variables on the fly, but again are beyond the purview of this guide.)

After running many multivariate models estat summarize will give you simple descriptive statistics for the specific sample used in that particular analysis.

Graphing

histogram x will give you a nice display of one variable. histogram x, by(y) may be useful for comparing the distributions of two variables over the categories of y.

histogram x, percent will scale the y-axis more intuitively in terms of percentages. histogram x, discrete gives a nicer display for categorical variables.

The percent and discrete options can be combined.

twoway scatter y x gives you a twoway scatterplot of your data. twoway lfit y x will give you a linear fit graph. The two syntaxes may be combined e.g. twoway (scatter y x)(lfit y x).

graph bar x, over(y) is useful for creating a bar graph of a continuous or categorical variable graphed across the categories of a categorical variable.

For all graphs, options after a “,” will be helpful in titling your graph e.g. twoway lfit y x, title(“.”) xtitle(“.”) ytitle(“.”)

by:

In many cases you may want to look at the results of some calculation for x, or x and y over a third variable z. In such cases the by: syntax will be especially useful. For example to look at the correlation of x and y over different values of z.

sort z by z: pwcorr x y, sig

Comments, questions and corrections most welcome and may be sent to: Andrew Grogan-Kaylor @ agrogan@umich.edu. This document available on the web @ https://agroganweb.wordpress.com/stata-resources/