 |
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
st: RE: heckprob dropping observations
<>
I cannot replicate the problem with the official example. What you can do is
single out the offending observations as in the following code and watch out
for any unusual patterns:
*******
webuse school, clear
replace vote=. in 2/4
heckprob private years logptax, sel(vote=years loginc logptax) nolog
l if !e(sample)
ins vote private if !e(sample)
su if !e(sample)
*******
HTH
Martin
-----Original Message-----
From: [email protected]
[mailto:[email protected]] On Behalf Of Carlena Ficano
Sent: Samstag, 13. Februar 2010 00:45
To: [email protected]
Subject: st: heckprob dropping observations
I am running a heckprob model (code and results pasted below), and the
routine appears to be dropping 2 observations from the dataset for reasons
that I cannot trace. I have been pouring through my data to see where I
might have missing values generating the problem, but I find none in the
first stage.
Specifically, I have 890 complete observations for all variables from the
first stage of the equation ("ret_12") and 785 complete observations for all
variables from the second stage of the equation ("ret_23"). To check this,
I ran a simple probit of the first stage and indeed all 890 observations
were included (again, code and results pasted below).
Has anyone else ever encountered such a problem, or am I missing something
simple?
Thanks-
Carli
Code for heckprob:
heckprob ret_23
hs_gradrate_2 missingrate_2
apunder3 pubhs_2 micro smtown rural ismale_1 white_1
busa_2 socsci_2 sci_2 nurs_2 unde_2 tot_cred_2 avadvmax avarat_2 avaratsq_2
athl_12r kav_cls_ar_2
need_2 aid_2
gpa2hat,
sel(ret_12=
hs_gradrate_1 missingrate_1
apunder3 pubhs_1 micro smtown rural ismale_1 white_1
busa_1 socsci_1 sci_1 nurs_1 unde_1 tot_cred_1 advmax_1 arat_1 aratsq_1
athl_1r kav_cls_ar_1
need_1 aid_1
gpa1hat) noskip;
Results:
Probit model with sample selection Number of obs =
888
Censored obs =
103
Uncensored obs =
785
LR chi2(23) =
62.46
Log likelihood = -584.4424 Prob > chi2 =
0.0000
----------------------------------------------------------------------------
--
| Coef. Std. Err. z P>|z| [95% Conf.
Interval]
-------------+--------------------------------------------------------------
--
ret_23 |
hs_gradrat~2 | .0013958 .0034865 0.40 0.689 -.0054377
.0082292
missingrat~2 | .2427822 .2236769 1.09 0.278 -.1956166
.6811809
apunder3 | .4432644 .1936606 2.29 0.022 .0636965
.8228322
pubhs_2 | .0973154 .1692737 0.57 0.565 -.2344551
.4290858
micro | -.0359948 .2012836 -0.18 0.858 -.4305033
.3585138
smtown | .5880747 .2843385 2.07 0.039 .0307814
1.145368
rural | -.0904672 .2146995 -0.42 0.673 -.5112705
.3303361
ismale_1 | -.0947296 .1216172 -0.78 0.436 -.333095
.1436357
white_1 | -.1234044 .1165534 -1.06 0.290 -.3518449
.1050361
busa_2 | .4007369 .2415171 1.66 0.097 -.0726279
.8741016
socsci_2 | .3898824 .1915603 2.04 0.042 .0144311
.7653338
sci_2 | -.0443619 .1999683 -0.22 0.824 -.4362925
.3475686
nurs_2 | .581178 .2613915 2.22 0.026 .0688602
1.093496
unde_2 | .2819771 .1740002 1.62 0.105 -.059057
.6230112
tot_cred_2 | .167432 .0357239 4.69 0.000 .0974146
.2374495
avadvmax | -.0090509 .0055766 -1.62 0.105 -.0199808
.0018791
avarat_2 | 3.687271 9.059758 0.41 0.684 -14.06953
21.44407
avaratsq_2 | -.4533746 1.061097 -0.43 0.669 -2.533087
1.626338
athl_12r | .0679246 .1154551 0.59 0.556 -.1583634
.2942125
kav_cls_ar_2 | .7582222 .2505294 3.03 0.002 .2671936
1.249251
need_2 | -5.76e-06 5.16e-06 -1.12 0.264 -.0000159
4.35e-06
aid_2 | 4.89e-06 .0000111 0.44 0.659 -.0000168
.0000266
gpa2hat | -.3899752 .098075 -3.98 0.000 -.5821986
-.1977518
_cons | -9.94018 19.36282 -0.51 0.608 -47.89061
28.01025
-------------+--------------------------------------------------------------
--
ret_12 |
hs_gradrat~1 | -.0025337 .0040619 -0.62 0.533 -.0104949
.0054274
missingrat~1 | -.2124467 .204638 -1.04 0.299 -.6135298
.1886365
apunder3 | -.1161432 .1888857 -0.61 0.539 -.4863524
.2540661
pubhs_1 | -.1429733 .1801124 -0.79 0.427 -.4959872
.2100405
micro | .4521774 .3854538 1.17 0.241 -.3032982
1.207653
smtown | -.3306497 .277702 -1.19 0.234 -.8749357
.2136363
rural | -.2278209 .2328102 -0.98 0.328 -.6841205
.2284786
ismale_1 | .3499555 .1462037 2.39 0.017 .0634015
.6365096
white_1 | -.2211232 .1341752 -1.65 0.099 -.4841017
.0418553
busa_1 | .3565705 .291409 1.22 0.221 -.2145807
.9277217
socsci_1 | .2729609 .232605 1.17 0.241 -.1829366
.7288584
sci_1 | .6075374 .2751517 2.21 0.027 .0682499
1.146825
nurs_1 | .5158331 .3119982 1.65 0.098 -.0956722
1.127338
unde_1 | .3032148 .2068252 1.47 0.143 -.1021552
.7085848
tot_cred_1 | .0769393 .0537845 1.43 0.153 -.0284765
.1823551
advmax_1 | -.0087446 .0061637 -1.42 0.156 -.0208252
.0033359
arat_1 | .8722621 5.456169 0.16 0.873 -9.821632
11.56616
aratsq_1 | -.1188429 .6444767 -0.18 0.854 -1.381994
1.144308
athl_1r | -.0742483 .1556409 -0.48 0.633 -.3792989
.2308022
kav_cls_ar_1 | -.8467959 .323886 -2.61 0.009 -1.481601
-.2119909
need_1 | 6.63e-06 6.60e-06 1.00 0.316 -6.32e-06
.0000196
aid_1 | -.0000228 .0000153 -1.49 0.136 -.0000528
7.18e-06
gpa1hat | 1.723617 .2469627 6.98 0.000 1.239579
2.207655
_cons | -3.197027 11.50147 -0.28 0.781 -25.7395
19.34545
-------------+--------------------------------------------------------------
--
/athrho | -4.009089 292.7936 -0.01 0.989 -577.874
569.8558
-------------+--------------------------------------------------------------
--
rho | -.9993414 .3855539 -1
1
----------------------------------------------------------------------------
--
LR test of indep. eqns. (rho = 0): chi2(1) = 6.86 Prob > chi2 =
0.0088
----------------------------------------------------------------------------
--
Probit code:
probit ret_12
hs_gradrate_1 missingrate_1
apunder3 pubhs_1 micro smtown rural ismale_1 white_1
busa_1 socsci_1 sci_1 nurs_1 unde_1 tot_cred_1 advmax_1 arat_1 aratsq_1
athl_1r kav_cls_ar_1
need_1 aid_1
gpa1hat;
Probit Results:
Probit regression Number of obs =
890
LR chi2(23) =
115.45
Prob > chi2 =
0.0000
Log likelihood = -261.18831 Pseudo R2 =
0.1810
----------------------------------------------------------------------------
--
ret_12 | Coef. Std. Err. z P>|z| [95% Conf.
Interval]
-------------+--------------------------------------------------------------
--
hs_gradrat~1 | -.0043464 .0040219 -1.08 0.280 -.0122293
.0035365
missingrat~1 | -.2415427 .2027886 -1.19 0.234 -.6390011
.1559157
apunder3 | -.1353076 .1878775 -0.72 0.471 -.5035408
.2329256
pubhs_1 | -.1342291 .1792748 -0.75 0.454 -.4856012
.217143
micro | .5242819 .3781409 1.39 0.166 -.2168607
1.265425
smtown | -.3376971 .2787695 -1.21 0.226 -.8840753
.208681
rural | -.2488978 .2346477 -1.06 0.289 -.7087988
.2110033
ismale_1 | .3851058 .1473312 2.61 0.009 .0963421
.6738696
white_1 | -.1924949 .1349971 -1.43 0.154 -.4570844
.0720946
busa_1 | .3320956 .2899227 1.15 0.252 -.2361425
.9003337
socsci_1 | .2384587 .2298812 1.04 0.300 -.2121001
.6890176
sci_1 | .4794838 .2726063 1.76 0.079 -.0548147
1.013782
nurs_1 | .5350326 .3088444 1.73 0.083 -.0702912
1.140356
unde_1 | .243891 .2053581 1.19 0.235 -.1586035
.6463855
tot_cred_1 | .0735117 .050484 1.46 0.145 -.0254351
.1724584
advmax_1 | -.0086672 .0061495 -1.41 0.159 -.02072
.0033856
arat_1 | 1.192344 5.417345 0.22 0.826 -9.425458
11.81014
aratsq_1 | -.1542927 .6402407 -0.24 0.810 -1.409141
1.100556
athl_1r | -.0986405 .1573756 -0.63 0.531 -.407091
.20981
kav_cls_ar_1 | -.7927225 .3183139 -2.49 0.013 -1.416606
-.1688388
need_1 | 7.31e-06 6.73e-06 1.09 0.277 -5.88e-06
.0000205
aid_1 | -.0000236 .0000157 -1.51 0.132 -.0000543
7.13e-06
gpa1hat | 1.646721 .2453282 6.71 0.000 1.165886
2.127555
_cons | -3.663654 11.41562 -0.32 0.748 -26.03786
18.71055
----------------------------------------------------------------------------
--
Carlena K. Cochi Ficano, Ph.D.
Associate Professor of Economics
Hartwick College
229 Golisano Hall
Oneonta, NY 13820
607-431-4946
607-431-4405 (fax)
*
* 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/
*
* 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/