Stata Commands and Table of Bivariate tests

This page offers detail on basic Stata commands along with a schema detailing the appropriate test to use bivariate analysis.

Basic Stata commands:

 

Stata commands

The basic Stata commands we will use in this workshop are:



Summary Statistics: this command provides summary statistics for variables v1 v2 and v3 and provides further detail.

su v1 v2 v3, d



Histogram: this command provides a histogram for v1.

hist v1



Scatter plot: this command provides a scatterplot for v1 and v2

scatter v1 v2



Correlation: this command provides the correlation coefficient between v1 and v2, with significance

corr v1 v2 (or "corr v1 v2, cov" for covariance)



Regression: this commands generates a regression with y as a function of x

regr y x



T-test: this command provides a ttest to determine if the mean of v1 is statistically significant between the two categories presented in v2 (v1 must be a continuous variable and v2 must be a dichotomous, i.e. 2-category, variable)

ttest v1, by(v2)



ANOVA: This test provides you an ANOVA test with associated tables of values and the post hoc Bonferroni test results.

oneway continuous categorical, t b

 


Chi-Square: this commands provides a Chi-square test to determine if two categorical variables are independent of one another. The options provide you with expected values, the chi-square test and the contriburion of each cell to the chi-square value.

tab v1 v2, exp chi2 cchi2

 

Basic schema detailing the appropriate bivariate tests to be done by type of variable:

Remember the schema I have asked you to put in your notes:

Continuous

Cat(2)

Cat(3+)

Continuous

Correlation

t-test

ANOVA

Cat(2)

t-test

 

Chisquare

Cat(3+)

ANOVA