Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: st: Generalized lineal models with survey data
From
Paolina Medina <[email protected]>
To
[email protected]
Subject
Re: st: Generalized lineal models with survey data
Date
Tue, 27 Jul 2010 11:30:09 -0500
Thank you both, very much.
So this almost zero alpha, without a confidence interval can be taken
to indicate that there is no overdispersion in the model?
Here is my svyset statement and the complete output..
I am using 52 regressors (including the constant), i really dont know
how many are the design degrees of freedom... But in fact whenever i
take any of these regressors i get a lot of troubles with convergence
in the survey results (not concave or backed up) and i have to throw
away many other regressors to get convergence again.
Do you know anything i can do to fix this?
Thank you very much again in advance!!
svyset upm [weight=factor], strata(estrato)
(sampling weights assumed)
pweight: factor
VCE: linearized
Strata 1: estrato
SU 1: upm
FPC 1: <zero>
. svy: nbreg ncels resmay6 numradios nTVs tfijo tpaga luz ncompus
internet prim2 sec2 prepa2 normal2 tec2 pro2 m2 doc2 traba
> jadores e1 e2 e3 e4 e5 e6 e7 e8 e9 e10 e11 e12 e13 e14 e15 e16 e17 e18 e19 e20 e21 e22 e23 e24 e25 e26 e27 e28 e29 e30 e31
> e32 estrato1 estrato2 estrato3 estrato4, log
(running nbreg on estimation sample)
note: e3 dropped due to collinearity
note: estrato2 dropped due to collinearity
Fitting Poisson model:
Iteration 0: log pseudolikelihood = -32297663
Iteration 1: log pseudolikelihood = -32264546
Iteration 2: log pseudolikelihood = -32264459
Iteration 3: log pseudolikelihood = -32264459
Fitting constant-only model:
Iteration 0: log pseudolikelihood = -41335387
Iteration 1: log pseudolikelihood = -40428544
Iteration 2: log pseudolikelihood = -40419335
Iteration 3: log pseudolikelihood = -40418972
Iteration 4: log pseudolikelihood = -40418972
Fitting full model:
Iteration 0: log pseudolikelihood = -34707498
Iteration 1: log pseudolikelihood = -32843264
Iteration 2: log pseudolikelihood = -32387977
Iteration 3: log pseudolikelihood = -32296994
Iteration 4: log pseudolikelihood = -32272196
Iteration 5: log pseudolikelihood = -32265972
Iteration 6: log pseudolikelihood = -32264708
Iteration 7: log pseudolikelihood = -32264488
Iteration 8: log pseudolikelihood = -32264465
Iteration 9: log pseudolikelihood = -32264460
Iteration 10: log pseudolikelihood = -32264459
Iteration 11: log pseudolikelihood = -32264459 (not concave)
Negative binomial regression Number of obs = 6089
LR chi2(50) = 1.63e+07
Dispersion = mean Prob > chi2 = 0.0000
Log pseudolikelihood = -32264459 Pseudo R2 = 0.2017
------------------------------------------------------------------------------
ncels | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
resmay6 | -.1189499 .0003869 -307.48 0.000 -.1197081 -.1181917
numradios | .0516967 .0002443 211.65 0.000 .0512179 .0521754
nTVs | .1765134 .0001923 918.12 0.000 .1761366 .1768902
tfijo | -.1133193 .0004484 -252.70 0.000 -.1141982 -.1124403
tpaga | .1687625 .0004486 376.22 0.000 .1678833 .1696417
luz | .7604265 .0046124 164.87 0.000 .7513864 .7694665
ncompus | .1401202 .0003325 421.45 0.000 .1394686 .1407718
internet | -.0047512 .0005739 -8.28 0.000 -.0058761 -.0036264
prim2 | .1220793 .0004182 291.90 0.000 .1212596 .122899
sec2 | .1243138 .0002344 530.28 0.000 .1238543 .1247733
prepa2 | .0985697 .0002857 344.95 0.000 .0980097 .0991298
normal2 | -.0754922 .0007591 -99.45 0.000 -.07698 -.0740044
tec2 | .0036673 .0007941 4.62 0.000 .002111 .0052236
pro2 | .0751997 .0004025 186.83 0.000 .0744109 .0759886
m2 | .0292361 .0007304 40.03 0.000 .0278046 .0306677
doc2 | -.160645 .002035 -78.94 0.000 -.1646335 -.1566565
trabajadores | .1317086 .0001916 687.47 0.000 .1313331 .1320841
e1 | .0016872 .002384 0.71 0.479 -.0029854 .0063599
e2 | .0648906 .002022 32.09 0.000 .0609276 .0688537
e4 | -.0181868 .0026577 -6.84 0.000 -.0233957 -.0129778
e5 | -.1868008 .0021267 -87.83 0.000 -.1909691 -.1826324
e6 | .1466343 .0026388 55.57 0.000 .1414623 .1518062
e7 | -.3935254 .0021961 -179.20 0.000 -.3978296 -.3892212
e8 | -.1245181 .0020253 -61.48 0.000 -.1284877 -.1205485
e9 | -.2299045 .0019112 -120.29 0.000 -.2336504 -.2261585
e10 | -.0837978 .0022478 -37.28 0.000 -.0882033 -.0793922
e11 | -.3531027 .0020698 -170.60 0.000 -.3571595 -.349046
e12 | -.4600149 .00238 -193.28 0.000 -.4646797 -.4553501
e13 | -.3038531 .002392 -127.03 0.000 -.3085414 -.2991648
e14 | -.127427 .0019299 -66.03 0.000 -.1312096 -.1236444
e15 | -.3696336 .0019058 -193.95 0.000 -.3733689 -.3658982
e16 | -.001945 .0020393 -0.95 0.340 -.0059418 .0020519
e17 | -.1287652 .0023403 -55.02 0.000 -.1333522 -.1241783
e18 | -.1066276 .0026009 -41.00 0.000 -.1117253 -.1015299
e19 | -.1713326 .0019934 -85.95 0.000 -.1752395 -.1674257
e20 | -.3297286 .0023891 -138.02 0.000 -.3344111 -.3250461
e21 | -.2838858 .0020318 -139.72 0.000 -.287868 -.2799037
e22 | -.0790438 .0022371 -35.33 0.000 -.0834285 -.0746591
e23 | .1517953 .0022013 68.96 0.000 .1474808 .1561099
e24 | -.2631786 .0022394 -117.52 0.000 -.2675677 -.2587895
e25 | .0587088 .0021346 27.50 0.000 .0545251 .0628925
e26 | .0442961 .0020797 21.30 0.000 .0402199 .0483723
e27 | .165134 .0021816 75.69 0.000 .1608581 .1694098
e28 | -.0281822 .0020233 -13.93 0.000 -.0321478 -.0242167
e29 | -.6082309 .002915 -208.66 0.000 -.6139442 -.6025177
e30 | -.1442515 .0020102 -71.76 0.000 -.1481914 -.1403116
e31 | .0250986 .0022023 11.40 0.000 .0207822 .0294149
e32 | -.1352461 .0025166 -53.74 0.000 -.1401785 -.1303137
estrato1 | .189495 .0005579 339.69 0.000 .1884016 .1905883
estrato3 | -.322916 .0010167 -317.62 0.000 -.3249087 -.3209234
estrato4 | -.479242 .0008096 -591.98 0.000 -.4808287 -.4776553
_cons | -1.493952 .0049766 -300.19 0.000 -1.503706 -1.484198
-------------+----------------------------------------------------------------
/lnalpha | -23.93108 . . .
-------------+----------------------------------------------------------------
alpha | 4.04e-11 . . .
------------------------------------------------------------------------------
Likelihood-ratio test of alpha=0: chibar2(01) = 0.00 Prob>=chibar2 = 1.000
Computing scores...
Survey results:
Survey: Negative binomial regression
Number of strata = 4 Number of obs = 6089
Number of PSUs = 837 Population size = 27782772
Design df = 833
F( 51, 783) = 50.61
Prob > F = 0.0000
------------------------------------------------------------------------------
| Linearized
ncels | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
resmay6 | -.1189499 .0302713 -3.93 0.000 -.1783669 -.0595329
numradios | .0516967 .019889 2.60 0.010 .0126581 .0907352
nTVs | .1765134 .0210548 8.38 0.000 .1351867 .2178402
tfijo | -.1133193 .0389423 -2.91 0.004 -.1897558 -.0368827
tpaga | .1687625 .0330961 5.10 0.000 .1038009 .233724
luz | .7604265 .4341484 1.75 0.080 -.091727 1.61258
ncompus | .1401202 .0232643 6.02 0.000 .0944567 .1857837
internet | -.0047512 .0434178 -0.11 0.913 -.0899725 .08047
prim2 | .1220793 .0353386 3.45 0.001 .0527162 .1914425
sec2 | .1243138 .0272578 4.56 0.000 .0708117 .1778159
prepa2 | .0985697 .0253691 3.89 0.000 .0487748 .1483647
normal2 | -.0754922 .0678928 -1.11 0.266 -.2087533 .0577689
tec2 | .0036673 .0752994 0.05 0.961 -.1441316 .1514662
pro2 | .0751997 .038429 1.96 0.051 -.0002293 .1506288
m2 | .0292361 .04043 0.72 0.470 -.0501205 .1085927
doc2 | -.160645 .0938375 -1.71 0.087 -.3448307 .0235407
trabajadores | .1317086 .0193343 6.81 0.000 .093759 .1696583
e1 | .0016872 .1667689 0.01 0.992 -.3256494 .3290239
e2 | .0648906 .1476771 0.44 0.660 -.2249723 .3547535
e4 | -.0181868 .1666332 -0.11 0.913 -.345257 .3088835
e5 | -.1868008 .1583675 -1.18 0.239 -.4976471 .1240456
e6 | .1466343 .1557808 0.94 0.347 -.1591347 .4524032
e7 | -.3935254 .2460117 -1.60 0.110 -.876401 .0893502
e8 | -.1245181 .1549826 -0.80 0.422 -.4287205 .1796842
e9 | -.2299045 .1487182 -1.55 0.123 -.5218109 .062002
e10 | -.0837978 .2982255 -0.28 0.779 -.6691594 .5015639
e11 | -.3531027 .1536798 -2.30 0.022 -.6547479 -.0514576
e12 | -.4600149 .1721165 -2.67 0.008 -.7978478 -.122182
e13 | -.3038531 .1661254 -1.83 0.068 -.6299266 .0222204
e14 | -.127427 .1614396 -0.79 0.430 -.4443031 .1894491
e15 | -.3696336 .151954 -2.43 0.015 -.6678914 -.0713758
e16 | -.001945 .2189627 -0.01 0.993 -.4317284 .4278385
e17 | -.1287652 .1711779 -0.75 0.452 -.4647559 .2072255
e18 | -.1066276 .162161 -0.66 0.511 -.4249198 .2116646
e19 | -.1713326 .1540608 -1.11 0.266 -.4737256 .1310603
e20 | -.3297286 .2472161 -1.33 0.183 -.8149682 .155511
e21 | -.2838858 .1542873 -1.84 0.066 -.5867234 .0189518
e22 | -.0790438 .1822687 -0.43 0.665 -.4368038 .2787162
e23 | .1517953 .1548563 0.98 0.327 -.1521591 .4557498
e24 | -.2631786 .179553 -1.47 0.143 -.6156081 .0892509
e25 | .0587088 .176449 0.33 0.739 -.2876282 .4050458
e26 | .0442961 .1611852 0.27 0.784 -.2720808 .3606729
e27 | .165134 .168757 0.98 0.328 -.1661049 .4963729
e28 | -.0281822 .1619115 -0.17 0.862 -.3459847 .2896203
e29 | -.6082309 .1689908 -3.60 0.000 -.9399288 -.2765331
e30 | -.1442515 .190234 -0.76 0.448 -.5176458 .2291428
e31 | .0250986 .1760391 0.14 0.887 -.3204338 .3706309
e32 | -.1352461 .2260741 -0.60 0.550 -.578988 .3084957
estrato1 | .189495 .0642525 2.95 0.003 .0633791 .3156108
estrato3 | -.322916 .111296 -2.90 0.004 -.5413697 -.1044624
estrato4 | -.479242 .1214264 -3.95 0.000 -.7175798 -.2409043
_cons | -1.493952 .4622372 -3.23 0.001 -2.401239 -.5866658
-------------+----------------------------------------------------------------
/lnalpha | -23.93108 . . .
-------------+----------------------------------------------------------------
alpha | 4.04e-11 . . .
------------------------------------------------------------------------------
On Tue, Jul 27, 2010 at 11:14 AM, Stas Kolenikov <[email protected]> wrote:
> Most likely, you run out of degrees of freedom. They should be
> reported in -svy:- output, and if you have more regressors/model
> parameters than the design d.f., you will have missing standard
> errors.
>
> On Tue, Jul 27, 2010 at 4:51 PM, Paolina Medina
> <[email protected]> wrote:
>> Thank you very much Steve for your help, in fact i have tried to run
>> the svy: nbreg command, but i am encountering some problems, i hope
>> you can help me:
>>
>> This is the regression that i run:
>>
>> svy: nbreg ncels residents numradios nTVs dfixedphone delectricity
>> ncompus dinternet elementary highschool phd workingpeople e1 e2 e3 e4
>> e5 e6 e7 e8 e9 e10 e11 e12 e13 e14 e15 e16 e17 e18 e19 e20 e21 e22 e23
>> e24 e25 e26 e27 e28 e29 e30 e31 e32 estrato1 estrato2 estrato3
>> estrato4, log
>>
>> I am trying to estimate coefficients for the number of cell phones in
>> mexican households using a technology specialized survey. I use some
>> socioeconomic variables, 32 dummies one for each state in mexico, and
>> 4 dummies indicating the size of the town.
>>
>> But in the first stage of the regresion, below the coefficients estimates i get:
>>
>> -------------+----------------------------------------------------------------
>> /lnalpha | -23.93108 . . .
>> -------------+----------------------------------------------------------------
>> alpha | 4.04e-11 . . .
>> ------------------------------------------------------------------------------
>> Likelihood-ratio test of alpha=0: chibar2(01) = 0.00 Prob>=chibar2 = 1.000
>>
>>
>> And in the survey results, below the coefficients estimates i get
>>
>> -------------+----------------------------------------------------------------
>> /lnalpha | -23.93108 . . .
>> -------------+----------------------------------------------------------------
>> alpha | 4.04e-11 . . .
>> ------------------------------------------------------------------------------
>>
>> I understand that stata wont calculate the LR test for alpha when it
>> comes to survey data, but as you may see it is not even giving me the
>> confidence interval!
>>
>> Do you happen to know what is happening?
>>
>> Thank you very very much in advance!
>>
>> Regards!
>>
>> PM
>>
>>
>>
>>
>>
>>
>>
>> On Mon, Jul 26, 2010 at 10:24 PM, Steve Samuels <[email protected]> wrote:
>>> In Version 9, -svy: nbreg- and -svy: gnbreg- will work for you. Both
>>> fit generalizations of the Poisson with extra dispersion. -svy: glm-
>>> was not available until Stata 10.
>>>
>>> Steve
>>>
>>>
>>> On Mon, Jul 26, 2010 at 10:00 PM, Paolina Medina
>>> <[email protected]> wrote:
>>>> Dear statalisters,
>>>> Is it possible to perform a regression using glm with survey data in stata 9?
>>>> I have count data with overdispersion (mean 1.21, variance 1.70).
>>>> Thank you very much in advance,
>>>> PM
>>>> *
>>>> * For searches and help try:
>>>> * http://www.stata.com/help.cgi?search
>>>> * http://www.stata.com/support/statalist/faq
>>>> * http://www.ats.ucla.edu/stat/stata/
>>>>
>>>
>>>
>>>
>>> --
>>> Steven Samuels
>>> [email protected]
>>> 18 Cantine's Island
>>> Saugerties NY 12477
>>> USA
>>> Voice: 845-246-0774
>>> Fax: 206-202-4783
>>>
>>> *
>>> * For searches and help try:
>>> * http://www.stata.com/help.cgi?search
>>> * http://www.stata.com/support/statalist/faq
>>> * http://www.ats.ucla.edu/stat/stata/
>>>
>>
>>
>>
>> --
>> Paolina Medina Palma
>>
>> *
>> * For searches and help try:
>> * http://www.stata.com/help.cgi?search
>> * http://www.stata.com/support/statalist/faq
>> * http://www.ats.ucla.edu/stat/stata/
>>
>
>
>
> --
> Stas Kolenikov, also found at http://stas.kolenikov.name
> Small print: I use this email account for mailing lists only.
>
> *
> * For searches and help try:
> * http://www.stata.com/help.cgi?search
> * http://www.stata.com/support/statalist/faq
> * http://www.ats.ucla.edu/stat/stata/
>
--
Paolina Medina Palma
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/