Hi Jeff and other Listers -
How about the question 2 of Robert? Below is the question:
>2 - Is it possible to estimate a one-way anova to detect group differences
> with svy commands in Stata 8 based on my above example?
In Robert's example, variable x has three categories: 0, 1, and 2. Is there any way to compare the three group differences concurrently (like a one-way anova) with svy commands?
Hong
----- Original Message -----
From: [email protected] (Jeff Pitblado, StataCorp LP)
Date: Tuesday, August 31, 2004 12:16 pm
Subject: Re: st: stumped with lincom after svymean
> "Robert Bozick" <[email protected]> asks about using -test-
> after-svymean, by()-:
>
> > Here is what I am trying to accomplish. I have a continuous
> variable y, and
> > a categorical variable x. The categorical variable x has three
> categories:
> > 0, 1, and 2. I want to test whether y is different among the
> categories of
> > my variable x. My data are collected using a stratfied, cluster
> design and
> > therefore I must use the survey commands.
> >
> > In previous versions of stata, I would use the code:
> >
> > svymean y, by(x)
> > svytest [x]0=[x]1
> >
> > Stata 8 says that svytest no longer exists and that I should use
> the lincom
> > command. Maybe the coffee hasn't set in, but I cannot figure
> out how to do
> > replicate the above code using the lincom command based on the
> information
> > from Stata's help screen.
> >
> > So my question is two-fold:
> > 1 - How do I test group differences using the svy command in
> Stata 8 based
> > on my above example?
> > 2 - Is it possible to estimate a one-way anova to detect group
> differences
> > with svy commands in Stata 8 based on my above example?
>
> Here is an example using the auto data (not that I remove the
> labels so that
> my example appears closer the one above):
>
> . sysuse auto, clear
> . label drop _all
> . svyset, srs
> . svymean mpg, by(foreign)
> . test [mpg]0 = [mpg]1
>
> When the -by()- option is supplied to -svymean-, the name of the
> variable that
> we are taking the mean of becomes something like an equation
> identifier when
> used with -test-. The values of the categories in the -by()-
> variable are
> used to identify which mean we are talking about.
>
> Here is the output from the above example:
>
> ***** BEGIN:
> . sysuse auto, clear
> (1978 Automobile Data)
>
>
> . label drop _all
>
>
> . svyset, srs
> simple random sample (SRS)
>
>
> . svymean mpg, by(foreign)
>
>
> Survey mean estimation
>
>
> pweight: <none> Number of obs
> = 74
> Strata: <one> Number of strata
> = 1
> PSU: <observations> Number of PSUs
> = 74
> Population size
> = 74
>
>
> -------------------------------------------------------------------
> -----------
> Mean Subpop. | Estimate Std. Err. [95% Conf. Interval]
> Deff
> ---------------+---------------------------------------------------
> -----------
> mpg |
> foreign==0 | 19.82692 .6558681 18.51978 21.13407
> 1
> foreign==1 | 24.77273 1.386503 22.00943 27.53602
> 1
> -------------------------------------------------------------------
> -----------
>
>
> . test [mpg]0 = [mpg]1
>
>
> Adjusted Wald test
>
>
> ( 1) [mpg]0 - [mpg]1 = 0
>
>
> F( 1, 73) = 10.40
> Prob > F = 0.0019
> ***** END:
>
> --Jeff
> [email protected]
> *
> * For searches and help try:
> * http://www.stata.com/support/faqs/res/findit.html
> * http://www.stata.com/support/statalist/faq
> * http://www.ats.ucla.edu/stat/stata/
>
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
