After using the command svyprop (or svytab for a one-way contingency
table), I would like to test whether the cell entries equal a vector of
values that I supply. It appears I may use svytest after svyprop, but I
fear the degrees of freedom will be wrong. For example, if I am
studying the response frequencies for a categorical variable with 5
possible values, and I supply a vector of 5 frequencies for the null
hypothesis, I suspect svytest will ignore the fact that I really only
have 4 freely varying relative frequencies.
One option would be to use svytest to compute the test statistic, and
simply correct the degrees of freedom myself. I could do this with a
simple chi-square test statistic, but I get lost in the design-corrected
F-tests used by the Stata svy procedures.
Two questions:
1) How may I conduct a hypothesis test with the correct degrees of
freedom, following svyprop or svytab?
2) What efficient syntax may I use to tell Stata my vector of null
hypothesis values? Must I write out p1=p1-hat, p2=p2-hat, ..., or can I
simply tell Stata the vector of null values?
I checked Stata manual and web resources, but couldn't find answers.
Thanks for your help!
Parke
Parke Wilde
USDA/Economic Research Service
1800 M St. NW, Room S2092
Washington DC 20036-5831
202-694-5633
*
* 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/
