A short guide to

commands
DR. MOHAMED ELSHERIF

1
Stata language
The standard Stata syntax
[by:] [command] [variablename(s)] [if] [in] [, options]

Variablename(s): can be one variable: [varname], or multiple variables [varlist] based on the command

If: to run the analysis only on a subset of the data

In: to run the analysis in selected of a range of observations (rows) in the dataset

Options: used to modify/add to the command

2
Stata language, helpful commands
by [varlist]
Prefix to run the commands separately in the categories of the specified variables. It comes after
sorting the data using that variable.
by [varlist]: [command] [variablename(s)] [if] [in] [, options]
by Sex: sum Age
sort [varlist]
Sorts the observations in the dataset from low to high based on the variables listed.
sort Sex

bysort [varlist]
It performs sorting and running the commands separately in one step (sort+by)
bysort Sex: sum Age

3
Stata language, helpful commands
[if] condition
to run the analysis only on a subset of the data based on a specific condition
Operators that can be used with if:

== Equal != Not equals

> Greater than & And
< Less than | Or
>= Greater than or equal to ! Not
<= Less than or equal to ~ Not
tab Sex if Age == 40
tab Sex if Age < 40

4
Stata language, helpful commands

display
To use Stata as a calculator.
display 2+3

help [command/text]
To get information about that command, or search for the text
help display

5
Exploring data
describe [varlist]

It gives name of the dataset, number of variables and observations, list of all variables with the
names, labels, and storage type.
describe
describe Age Drug
inspect [varlist]

It shows the number of missing and non-missing values and the number of unique values.
inspect
inspect Age Drug

6
Exploring data
codebook [varlist]

It shows the name, label, type, range, number of observations and number of missing values of
each variable in the dataset.
codebook
codebook Age Drug
codebook [varlist] , compact

This option shows the number of observations (obs), number of different values (unique),
mean, lowest value (min), highest value (max) and the variable label (label).
codebook , compact
codebook Age Drug , compact

7
Exploring data, listing observations

list [varlist] if

It shows a list of cases with data of all variables mentioned in [varlist] that meet the condition
list Height Weight Marital if Age == 35
list Height Weight Marital
list if Age == 35

8
Summarizing data, numeric variables
summarize [varlist]
It shows the number of observations, mean, SD and range.
sum
sum Age Height Weight
summarize [varlist] , detail
It gives extra details as percentiles, largest 4 values, smallest 4 values, and others.
sum , detail
sum Age Height Weight , detail
mean [varlist]
it gives the mean , standard error, and 95% CI of the mean.
mean Age
9
Summarizing data, numeric variables
tabstat [varlist] , statistics ( name of required statistics )

It is used to get specified summary statistics.

tabstat Age Height Weight , s( mean sd median iqr )
The option col(s) can be used for the summary statistics to be on columns instead of rows
tabstat Age Height Weight , s( mean sd median iqr ) col(s)
Some summary statistics that can be obtained from this command:

mean mean range range = max – min

sd standard deviation median median
max maximum iqr interquartile range = p75 - p25
min minimum cv coefficient of variation (sd/mean)

10
Summarizing data, numeric variables in
groups
bysort [varlist] : summarize [varlist]
It shows the summary statistics for each category of the variables listed after by or bysort.
bysort Sex: sum Age Height Weight

bysort [varlist] : tabstat [varlist] , statistics ( )

It shows the summary statistics for each category of the variables listed after by or bysort.
bysort Sex: tabstat Age Height Weight , s( mean sd median iqr ) col(s)

mean [varlist] , over ([varlist])

It gives the mean , standard error, and 95% CI of the mean in each of the specified
categories in the option over().
mean Age , over (Sex)
11
Summarizing data, categorical variables
tabulate [varname]
It shows the numbers (frequency) and percentages (relative frequency) for a specific variable
tab Sex
tab1 [varlist]
It shows the numbers (frequency) and percentages (relative frequency) for multiple
variables
tab1 Sex Edu Marital Excercise
bysort [varname]: tabulate [varname]
To get numbers and percentages for a specific variable in each of the categories of another variable
bysort Edu: tab Sex

12
Summarizing data, categorical variables
(two way)
tabulate [varname1] [varname2]
It gives a two-way table for the TOW specified variables
tab Sex Edu
tabulate [varname1] [varname2] , row

It gives a two-way table for the TOW specified variables with percentages per rows
tab Sex Edu , row

tabulate [varname1] [varname2] , column

It gives a two-way table for the TOW specified variables with percentages per columns
tab Sex Edu , col

13
Testing for normality of distribution
Shapiro-Wilk test for normality
swilk [varlist]
It is used for normality testing of specified variables
swilk Age Height

bysort [varlist] : swilk [varlist]

It is used for normality testing of specified variables in different categories
bysort Sex : swilk Age Height

14
Testing for equality of variance
F test for equality of variance (homogeneity of variance)
sdtest [varname] , by ([groupingvariable])
It is used for testing equality of variance between groups for a specified variable
sdtest Age , by (Sex)
Levene’s test for equality of variance (homogeneity of variance)
robvar [varname] , by ([groupingvariable])
It is used for testing equality of variance between groups for a specified variable
robvar Age , by (Sex)
Bartlett's test for equality of variance
oneway [varname] [groupingvariable]
It is used for testing equality of variance, and it is part of the oneway command output
oneway Age Sex
15
One-sample t-test
ttest [varname] =[value]

To test if the mean of a variable is equal to a specified value (for a single group)
ttest Age =35

16
Paired samples t test
ttest [varname1] = [varname2]

To test if the mean of the two variables is equal (difference is equal to zero). Data is presented in
pairs
ttest Exampre = Examafter

17
Independent samples t-test
With equality of variance
ttest [varname] , by ([groupingvariable])

To compare the mean of two groups (testing if the mean in the two groups is equal)
Before running the test, we need to check for the normality of distribution and equality of
variance.
ttest Age , by (Sex)

Without equality of variance

ttest [varname] , by ([groupingvariable]) unequal
If there is no equality of variance, the option unequal is added
ttest Age , by (Sex) unequal
18
One-way ANOVA
With equality of variance
oneway [varname] [groupingvariable]

To compare the mean of more than two groups (testing if the mean in the groups is equal)
Before running the test, we need to check for the normality of distribution and equality of
variance. Equality of variance is part of the output “Bartlett's test”
oneway Age Edu

To add table for summary statistics

oneway [varname] [groupingvariable] , tab
Option tab is used to get summary statistics as mean and sd per group
oneway Age Edu , tab
19
One-way ANOVA
To get the result for Bonferroni post hoc test
oneway [varname] [groupingvariable] , bonferroni

To get the result for Bonferroni post hoc test

oneway Age Edu , bonferroni

If there is no equality of variance, Welch ANOVA test is used

findit wtest
wtest [varname] [groupingvariable]

First, we need to install the wtest command using “findit wtest”

wtest Age Edu

20
Correlation
Pearson’s correlation
pwcorr [varlist]

To get the Pearson’s correlation coefficient between two numeric variables.

pwcorr Age Height
If more than two variables are used, a correlation matrix is produced
pwcorr Age Height Weight
To add the p-value and number of observations
pwcorr [varlist] , sig obs
Option sig is used to get p-value and option obs is used to get the number of observations (pairs)
pwcorr Age Height Weight , sig obs
21
Correlation
Confidence interval for Pearson’s correlation coefficient
findit corrci
corrci [varlist]
To get the Pearson’s correlation coefficient with the 95% CI. The command corrci needs to be
installed first using “findit corrci”
corrci Age Height Weight

Spearman’s correlation
spearman [varlist]
To get the Spearman’s correlation coefficient between two variables.
spearman Age Height Weight
22
Mann Whitney test =
Two-sample Wilcoxon rank-sum
ranksum [varname] , by ([groupingvariable])

It is the non-parametric equivalent of independent samples t-test to compare two groups

ranksum Age , by (Sex)

23
Wilcoxon signed-rank test

signrank [varname1] = [varname2]

It is the non-parametric equivalent of paired samples t-test to compare two variables (paired
data)
signrank Exampre = Examafter

24
Kruskal Wallis test
Kruskal Wallis test
kwallis [varname] , by ([groupingvariable])

It is the non-parametric equivalent of one-way ANOVA to compare more than two groups
kwallis Age , by (Edu)

25
Chi-square test
Chi-square test
tabulate [varname1] [varname2] , chi2

It is the non-parametric test to study the association between two categorical variables
tab Sex Edu , chi2
Chi-square test, with percentages per row
tabulate [varname1] [varname2] , chi2 row

tab Sex Edu , chi2 row

Chi-square test, with percentages per column

tabulate [varname1] [varname2] , chi2 col

tab Sex Edu , chi2 col

26
Chi-square test
Chi-square test, with expected values
tabulate [varname1] [varname2] , chi2 expect

tab Sex Edu , chi2 expect

Chi-square test, with other measures of association

tabulate [varname1] [varname2] , all

tab Sex Edu , all

Fisher’s exact test

tabulate [varname1] [varname2] , exact

tab Sex Edu , exact

27
Linear regression
Linear regression
regress [dependent var] [independent varlist]

regress Weight Age Height

Linear regression using categorical independent variable

regress [dependent var] i.[independent variable]

We add i. before the independent categorical variable

the first category is taken by default as reference category
regress Weight Age Height i.Edu

28
Logistic regression
Logistic regression
logit [dependent var] [independent varlist]
logit StatsTraining Age Height Weight
Logistic regression reporting OR
logit [dependent var] [independent varlist] , or

logistic [dependent var] [independent varlist]

To report the odds ratio, we use the option , or
Another option is to use the command logistic
logit StatsTraining Age Height Weight , or
logistic StatsTraining Age Height Weight

29