In the first and third examples it looks as though confint(svymean())
matches the totals and svyciprop(method="logit") matches the
proportions, which is what Stata says
(http://www.stata.com/statalist/archive/2006-10/msg01127.html). The
agreement isn't perfect for the counts in the first example,
Hi Dr. Lumley, you're obviously correct about all of that. Thank you for
cluing me into it! And sorry for overlooking that part of the
documentation.
I'm unfortunately still struggling with matching numbers exactly, and I
foolishly provided a dataset without a weight variable - thinking there wa
On Sat, Sep 22, 2012 at 2:51 AM, Anthony Damico wrote:
> Survey: Mean estimation
>
> Number of strata = 1 Number of obs= 183
> Number of PSUs = 15 Population size = 9235.4
> Design df= 14
>
> ---
Hi everyone, apologies if the answer to this is in an obvious place. I've
been searching for about a day and haven't found anything..
I'm trying to replicate Stata's confidence intervals in R with the survey
package, and the numbers are very very close but not exact. My ultimate
goal is to repl
4 matches
Mail list logo
