Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, is already up and running.

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: RE: test command for categorical variables in survey data

From   Steve Samuels <>
Subject   Re: st: RE: test command for categorical variables in survey data
Date   Tue, 28 Feb 2012 18:04:37 -0500

Nalesta Pilgrim:

The test in -svy: tab- (like the chi square test in -tabulate twoway- is a test of independence, not a test of equality of proportions. (And, with -svy-, the distribution of the test is not Chi Square, but that is a different matter.)
You are apparently mixing up the chi square test with the test for equality of means in one-way ANOVA. It's more analogous to a test for interactions in ANOVA.

We need an example of what you want to do. Start with this code and point out proportions you wish to test for equality. 

webuse nhanes2b
svy: tabulate race diabetes


On Feb 23, 2012, at 2:01 PM, Nanlesta Pilgrim wrote:

Dear Statalist,
I'm wondering if there is a test command for categorical variables
with survey data.
With continuous variables, you can run the test command:
. svy: mean testresult, over(childage)
. test [testresult]1=[testresult]2

For categorical variables, you run:
svy: tab wealth province

My question: how can I test which wealth category is significantly
different from another wealth category if the overall test result
indicates there is a significant difference but all the confidence
intervals overlap?

*   For searches and help try:

*   For searches and help try:

© Copyright 1996–2015 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index