|
|  |
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
st: xtlogit and constant term
Deal all
I am regressing the same model w/ and w/out the intercept using "xtlogit, re" in an unbalanced
panel data. However the odds' signs change if the constant is dropped.
Could you help me explain this? (I am copying both outputs)
I appreciate you help
Juan Julio Gutierrez
PhD Student
George Mason University
School Of Public Policy
MODEL 1 NO CONSTANT
. xtlogit DO2 INDICATOR_1 INDICATOR_2 INDICATOR3r INDICATOR4 INDICATOR5 INDICATOR6 INDICATOR7
INDICATOR860 NINDICATOR5 NINDICATOR6 ratif3 NINDICATOR9IP NINDICATOR10 if time>3, re i (project)
nocons
Fitting comparison model:
Iteration 0: log likelihood = -3112.924
Iteration 1: log likelihood = -1638.9574
Iteration 2: log likelihood = -1439.5861
Iteration 3: log likelihood = -1409.2021
Iteration 4: log likelihood = -1407.888
Iteration 5: log likelihood = -1407.8844
Fitting full model:
tau = 0.0 log likelihood = -1247.9671
tau = 0.1 log likelihood = -1200.3429
tau = 0.2 log likelihood = -1154.7556
tau = 0.3 log likelihood = -1111.6765
tau = 0.4 log likelihood = -1071.3706
tau = 0.5 log likelihood = -1034.3332
tau = 0.6 log likelihood = -1001.3773
tau = 0.7 log likelihood = -973.9988
tau = 0.8 log likelihood = -955.79642
Iteration 0: log likelihood = -1097.1048
Iteration 1: log likelihood = -1038.5596
Iteration 2: log likelihood = -1001.1093
Iteration 3: log likelihood = -999.11373
Iteration 4: log likelihood = -999.11102
Iteration 5: log likelihood = -999.11102
Random-effects logistic regression Number of obs = 4491
Group variable (i): project Number of groups = 727
Random effects u_i ~ Gaussian Obs per group: min = 1
avg = 6.2
max = 15
Wald chi2(13) = 767.98
Log likelihood = -999.11102 Prob > chi2 = 0.0000
------------------------------------------------------------------------------
DO2 | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
INDICATOR_1 | -.8357329 .239958 -3.48 0.000 -1.306042 -.3654238
INDICATOR_2 | -.4110548 .2116675 -1.94 0.052 -.8259155 .003806
INDICATOR3r | -.2578675 .3267794 -0.79 0.430 -.8983434 .3826083
INDICATOR4 | -1.469595 .1437704 -10.22 0.000 -1.75138 -1.18781
INDICATOR5 | 1.997214 .604792 3.30 0.001 .8118435 3.182585
INDICATOR6 | -3.033746 .5477888 -5.54 0.000 -4.107392 -1.9601
INDICATOR7 | -.8611347 .208806 -4.12 0.000 -1.270387 -.4518824
INDICATOR860 | -1.39018 .2136582 -6.51 0.000 -1.808942 -.9714177
NINDICATOR5 | -.478409 .1795451 -2.66 0.008 -.8303109 -.1265072
NINDICATOR6 | -.651115 .5510574 -1.18 0.237 -1.731168 .4289378
ratif3 | -3.337644 .2498051 -13.36 0.000 -3.827253 -2.848035
NINDICATOR~P | 3.350268 .1965227 17.05 0.000 2.965091 3.735445
NINDICATOR10 | 4.86e-10 4.64e-10 1.05 0.295 -4.23e-10 1.40e-09
-------------+----------------------------------------------------------------
/lnsig2u | 1.832401 .0796272 1.676334 1.988467
-------------+----------------------------------------------------------------
sigma_u | 2.499774 .099525 2.312125 2.702652
rho | .6551045 .0179912 .6190427 .689465
------------------------------------------------------------------------------
Likelihood-ratio test of rho=0: chibar2(01) = 817.55 Prob >= chibar2 = 0.000
MODEL 2 W/ CONSTANT
. xtlogit DO2 INDICATOR_1 INDICATOR_2 INDICATOR3r INDICATOR45 INDICATOR5 INDICATOR6 INDICATOR7
INDICATOR860 NINDICATOR5 NINDICATOR6 ratif3 NINDICATOR9IP NINDICATOR10 if time>3, re i (project)
Fitting comparison model:
Iteration 0: log likelihood = -1189.2225
Iteration 1: log likelihood = -843.65744
Iteration 2: log likelihood = -776.35165
Iteration 3: log likelihood = -766.59039
Iteration 4: log likelihood = -766.19648
Iteration 5: log likelihood = -766.19547
Fitting full model:
tau = 0.0 log likelihood = -766.19547
tau = 0.1 log likelihood = -742.84168
tau = 0.2 log likelihood = -721.61453
tau = 0.3 log likelihood = -702.20943
tau = 0.4 log likelihood = -685.01682
tau = 0.5 log likelihood = -670.85859
tau = 0.6 log likelihood = -660.99182
tau = 0.7 log likelihood = -658.22243
tau = 0.8 log likelihood = -668.02088
Iteration 0: log likelihood = -658.44381
Iteration 1: log likelihood = -614.32708
Iteration 2: log likelihood = -612.76708
Iteration 3: log likelihood = -612.76061
Iteration 4: log likelihood = -612.76061
Random-effects logistic regression Number of obs = 4491
Group variable (i): project Number of groups = 727
Random effects u_i ~ Gaussian Obs per group: min = 1
avg = 6.2
max = 15
Wald chi2(13) = 382.86
Log likelihood = -612.76061 Prob > chi2 = 0.0000
------------------------------------------------------------------------------
DO2 | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
INDICATOR_1 | .754484 .2759757 2.73 0.006 .2135816 1.295386
INDICATOR_2 | .7503506 .2508757 2.99 0.003 .2586433 1.242058
INDICATOR3r | .5662597 .3459642 1.64 0.102 -.1118176 1.244337
INDICATOR45 | .7463463 .4448553 1.68 0.093 -.1255541 1.618247
INDICATOR5 | 1.270672 .6709532 1.89 0.058 -.0443719 2.585716
INDICATOR6 | -.4390345 .5438313 -0.81 0.419 -1.504924 .6268552
INDICATOR7 | .0692061 .2361465 0.29 0.769 -.3936325 .5320446
INDICATOR860 | .6912012 .2532381 2.73 0.006 .1948635 1.187539
NINDICATOR5 | .6963986 .2112728 3.30 0.001 .2823116 1.110486
NINDICATOR6 | -.4586759 .6215281 -0.74 0.461 -1.676849 .7594968
ratif3 | .2452439 .2917346 0.84 0.401 -.3265454 .8170331
NINDICATOR~P | 4.428921 .2416649 18.33 0.000 3.955266 4.902575
NINDICATOR10 | 2.24e-10 5.46e-10 0.41 0.682 -8.46e-10 1.29e-09
_cons | -7.131021 .3294699 -21.64 0.000 -7.77677 -6.485272
-------------+----------------------------------------------------------------
/lnsig2u | 1.420105 .1020098 1.220169 1.62004
-------------+----------------------------------------------------------------
sigma_u | 2.034098 .1037489 1.840587 2.247953
rho | .5570646 .0251703 .5073299 .6056809
------------------------------------------------------------------------------
Likelihood-ratio test of rho=0: chibar2(01) = 306.87 Prob >= chibar2 = 0.000
__________________________________________________
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
http://mail.yahoo.com
*
* 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/
