Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: RE: hypothesis tests after svyprop


From   "Nick Winter" <nwinter@policystudies.com>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: RE: hypothesis tests after svyprop
Date   Thu, 17 Oct 2002 11:45:50 -0400

-svyprop- is not an estimation command; it does not return anything in
e(b) and e(V), so no post-estimation testing is possible.

-svytab- does return its results in e(b).  For a one-way table, the
columns will be labeled "p11" "p21" "p31" ... and so on.  So you can use
-svytest- to do a joint test all but one of those equal whatever values
you are interested in.  The only trick is that -svytest- requires that
there be variables named "p11" "p21", etc., although it does not
actually look at the contents of those variables to do the test.  So
this code would do it:

. svytab age one
. gen p11=.
. gen p21=.
. gen p31=.
. gen p41=.
. gen p51=.

. svytest p11=.19, notest
. svytest p21=.19, accum notest
. svytest p31=.19, accum notest
. svytest p41=.19, accum notest

Note that you test all but one of the proportions; the fifth is
redundant (and would be dropped by Stata anyway).  You'll get the same
answer regardless of the four you pick, as long as your null proportions
add up to 1.

Whether a Wald test is appropriate for this, or the best available, I
haven't thought much about. 

--Nick Winter

*
*   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/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index