 |
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
Re: st: Problems with standard errors using xi3, svy and ologit
At 01:37 PM 2/14/2008, Iain Lang wrote:
Dear Statalist members
I'm running models in Stata 9 using xi3, svy and ologit. This
combination seems to produce very large standard errors, and these
don't occur when I use xi instead of xi3. An example is pasted below
(using a binary outcome as an example). Has anyone else encountered
similar problems? Any help or comments appreciated!
With regards
Iain
I can't explain it, but I can confirm it:
. webuse nhanes2f
. xi3: svy: ologit health i.sex
i.sex _Isex_1-2 (naturally coded; _Isex_1 omitted)
(running ologit on estimation sample)
Survey: Ordered logistic regression
Number of strata = 31 Number of obs = 10335
Number of PSUs = 62 Population size = 1.170e+08
Design df = 31
F( 1, 31) = 6.79
Prob > F = 0.0140
------------------------------------------------------------------------------
| Linearized
| Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
health |
_Isex_2 | -.1917675 .0735911 -2.61 0.014 -.3418575 -.0416775
-------------+----------------------------------------------------------------
cut1 |
_cons | -3.106657 .0273867 -113.44 0.000 -3.162512 -3.050801
-------------+----------------------------------------------------------------
cut2 |
_cons | -1.68755 .077873 -21.67 0.000 -1.846373 -1.528727
-------------+----------------------------------------------------------------
cut3 |
_cons | -.3021831 .0652513 -4.63 0.000 -.435264 -.1691023
-------------+----------------------------------------------------------------
cut4 |
_cons | .8705437 .0547917 15.89 0.000 .7587954 .9822921
------------------------------------------------------------------------------
. xi: svy: ologit health i.sex
i.sex _Isex_1-2 (naturally coded; _Isex_1 omitted)
(running ologit on estimation sample)
Survey: Ordered logistic regression
Number of strata = 31 Number of obs = 10335
Number of PSUs = 62 Population size = 1.170e+08
Design df = 31
F( 1, 31) = 15.26
Prob > F = 0.0005
------------------------------------------------------------------------------
| Linearized
health | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
_Isex_2 | -.1917675 .0490878 -3.91 0.000 -.2918826 -.0916523
-------------+----------------------------------------------------------------
/cut1 | -3.106657 .0873699 -35.56 0.000 -3.284849 -2.928465
/cut2 | -1.68755 .0570402 -29.59 0.000 -1.803884 -1.571215
/cut3 | -.3021831 .0369428 -8.18 0.000 -.3775284 -.2268378
/cut4 | .8705437 .03337 26.09 0.000 .8024852 .9386022
------------------------------------------------------------------------------
It looks to me like it is related to version control. xi3 sets
version to 7. When I do the same, xi produces the same results that xi3 does:
. version 7: xi: svy: ologit health i.sex
i.sex _Isex_1-2 (naturally coded; _Isex_1 omitted)
(running ologit on estimation sample)
Survey: Ordered logistic regression
Number of strata = 31 Number of obs = 10335
Number of PSUs = 62 Population size = 1.170e+08
Design df = 31
F( 1, 31) = 6.79
Prob > F = 0.0140
------------------------------------------------------------------------------
| Linearized
| Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
health |
_Isex_2 | -.1917675 .0735911 -2.61 0.014 -.3418575 -.0416775
-------------+----------------------------------------------------------------
cut1 |
_cons | -3.106657 .0273867 -113.44 0.000 -3.162512 -3.050801
-------------+----------------------------------------------------------------
cut2 |
_cons | -1.68755 .077873 -21.67 0.000 -1.846373 -1.528727
-------------+----------------------------------------------------------------
cut3 |
_cons | -.3021831 .0652513 -4.63 0.000 -.435264 -.1691023
-------------+----------------------------------------------------------------
cut4 |
_cons | .8705437 .0547917 15.89 0.000 .7587954 .9822921
------------------------------------------------------------------------------
As to solutions - I suppose you could use xi3 to compute the vars
first, and then use the new vars in your command. You could also try
creating xi3b, change its version command, and see if it worked right then.
-------------------------------------------
Richard Williams, Notre Dame Dept of Sociology
OFFICE: (574)631-6668, (574)631-6463
HOME: (574)289-5227
EMAIL: [email protected]
WWW: http://www.nd.edu/~rwilliam
*
* 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/