# st: nestreg and survey logistic

 From Steve To statalist@hsphsun2.harvard.edu Subject st: nestreg and survey logistic Date Thu, 10 Nov 2005 15:55:53 -0600

Dear Stata Folk,

Stata exec and ado updated today on a WinXP system. Using Stata SE.

I've been trying out the new prefix command -nestreg-, repeating a sequence of analyses I've already done by hand.

The command syntax presents the two nested analyses correctly but then exits strangely with the message that -svy- was not found. The run is reproduced below:

. nestreg: svy: logistic past (store_water pluvial AC people vicinity cross3 aegypti missinginc inc100 intact) (city)

Block 1: store_water pluvial AC people vicinity cross3 aegypti missinginc inc100 intact
(running logistic on estimation sample)

Survey: Logistic regression

Number of strata = 2 Number of obs = 578
Number of PSUs = 100 Population size = 524546.74
Design df = 98
F( 10, 89) = 3.96
Prob > F = 0.0002

------------------------------------------------------------------------------
| Linearized
past | Odds Ratio Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
store_water | 1.462093 .405928 1.37 0.174 .842747 2.536604
pluvial | .5257532 .1184646 -2.85 0.005 .3361928 .8221962
AC | .6066694 .1344247 -2.26 0.026 .3908287 .9417112
people | 1.059671 .0605486 1.01 0.313 .9460759 1.186905
vicinity | 2.997936 .98036 3.36 0.001 1.566714 5.736608
cross3 | .8914012 .2043967 -0.50 0.617 .565528 1.405052
aegypti | .8925234 .2273284 -0.45 0.656 .5384005 1.479564
missinginc | .9671007 .2849121 -0.11 0.910 .5389754 1.7353
inc100 | 1.988252 .5067608 2.70 0.008 1.198966 3.297129
intact | 1.196331 .2396654 0.89 0.373 .8038874 1.780359
------------------------------------------------------------------------------

Block 2: city
(running logistic on estimation sample)

Survey: Logistic regression

Number of strata = 2 Number of obs = 578
Number of PSUs = 100 Population size = 524546.74
Design df = 98
F( 11, 88) = 6.70
Prob > F = 0.0000

------------------------------------------------------------------------------
| Linearized
past | Odds Ratio Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
store_water | .8597424 .2587414 -0.50 0.617 .4731455 1.562219
pluvial | .7104486 .1753382 -1.39 0.169 .4353418 1.159404
AC | .9446747 .232598 -0.23 0.818 .5795341 1.539875
people | 1.048103 .0607217 0.81 0.419 .9342719 1.175804
vicinity | 3.562469 1.236586 3.66 0.000 1.788927 7.094301
cross3 | .8968913 .2122723 -0.46 0.647 .5607413 1.434555
aegypti | .835552 .2144141 -0.70 0.486 .502126 1.390382
missinginc | .7654352 .2348365 -0.87 0.386 .4163838 1.407094
inc100 | 1.345265 .3731241 1.07 0.288 .775829 2.33265
intact | 1.115736 .2364086 0.52 0.606 .7327384 1.698923
city | 4.538661 1.255077 5.47 0.000 2.621816 7.856939
------------------------------------------------------------------------------

r(111);
_____

Since the Wald table is the default in the -nestreg- command, I was expecting the table to appear instead of the error message.

Have I made a simple syntax error or is there a bug in this new prefix command when using survey regressions?

Steve Rothenberg
CINVESTAV-IPN
Mérida, México

*
* 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/