# st: RE: Bivaraite versus univariate definition

 From "Nick Cox" To Subject st: RE: Bivaraite versus univariate definition Date Tue, 10 Feb 2009 21:33:31 -0000

```An analysis is bivariate if you need two variables to carry it out. It
is univariate if you only need one variable to carry out. That does not
stop you repeating such a method separately on several variables.

Chi-square statistics can arise in univariate, bivariate and
multivariate problems. There is no single thing uniquely called a
chi-square test.

There is no rule either way, but I'll assert from experience that using
a real name increases your chances of getting good answers on this list.

Nick
n.j.cox@durham.ac.uk

Anon Mouse

I have a very basic statistics question.  I have just started using
stata, so thank you for your help.

My question concerns the definition of bivariate versus univariate.

When is one analysis considered bivariate, and when is another
considered univariate?

I have seen univariate describe tests involving 1 predictor and 1 output
variable, but aI have also seen them described as basically descriptive
statistics.

I have seen bivaraite used to describe Chi2 tests.

So what would I call logistic regression analysis involving only one
predictor variable, e.g. high mpg (see example below)?  Univariate, or
bivariate analysis?

What would I call Chi2 test?  Univariate, or bivariate analysis?

***BEGIN EXAMPLE

sysuse auto

generate highmpg=1 if mpg>=20
replace highmpg=0 if mpg
_________________________________________________________________

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/
```