Dear forum members,
Currently I am running a linear regression (OLS - Difference in Difference setting) which includes multiple independent variables (continuous, binary, categorical, interaction terms). My regression includes one binary independent variable (BH=0/1) which depends on the state the observation stems from (7 out of 37 states --> BH=1, 30 out of 37 states --> BH=0). When running my estimation and including the categorical variable i.surveystate one state serves as baseline as it should but one more state gets omitted and I can not figure out why. When changing my binary BH variable (e.g. 4 out of 37 states --> BH=1) again one state is omitted in addition to the baseline state. So no matter how I define my treatment group BH always one state serves as baseline rightfully and another gets omitted (changing state depending on BH definition).
This issue does not occur when I define the binary BH based on another variable (issue of ommited state only occurs when BH is defined based on surveystate).
Could anyone help me with this issue even in general term (omitting one specification of categorical variable in addition to baseline although no observations are missing and this omitted specification changes).
Thank you very much in advance!
Greetings Caspar
Regression function
Example with Sokoto as baseline state (not depicted) and Kaduna as omitted state - BH=1 if observation from Borno, Yobe, Adamawa, Kano, Gombe, Bauchi, Kaduna state
Example with Sokoto as baseline state (not depicted) and Adamawa as omitted state - BH=1 if observation from Borno, Yobe, Adamawa state
Currently I am running a linear regression (OLS - Difference in Difference setting) which includes multiple independent variables (continuous, binary, categorical, interaction terms). My regression includes one binary independent variable (BH=0/1) which depends on the state the observation stems from (7 out of 37 states --> BH=1, 30 out of 37 states --> BH=0). When running my estimation and including the categorical variable i.surveystate one state serves as baseline as it should but one more state gets omitted and I can not figure out why. When changing my binary BH variable (e.g. 4 out of 37 states --> BH=1) again one state is omitted in addition to the baseline state. So no matter how I define my treatment group BH always one state serves as baseline rightfully and another gets omitted (changing state depending on BH definition).
This issue does not occur when I define the binary BH based on another variable (issue of ommited state only occurs when BH is defined based on surveystate).
Could anyone help me with this issue even in general term (omitting one specification of categorical variable in addition to baseline although no observations are missing and this omitted specification changes).
Thank you very much in advance!
Greetings Caspar
Regression function
Code:
svy: regress teenpreg i.BH i.post2009 i.post2009#i.BH i.muslim i.urban i.kanuri i.hhheadmale i.literacy i.wealthindex i.edulevel c.eduyearspartner i.largefamily i.surveystate
Example with Sokoto as baseline state (not depicted) and Kaduna as omitted state - BH=1 if observation from Borno, Yobe, Adamawa, Kano, Gombe, Bauchi, Kaduna state
Code:
surveystate |
zamfara | -.0458117 .0321642 -1.42 0.155 -.1088928 .0172693
katsina | .1005191 .0270401 3.72 0.000 .0474876 .1535506
jigawa | .0174331 .0300404 0.58 0.562 -.0414827 .0763489
yobe | -.0700439 .0300653 -2.33 0.020 -.1290086 -.0110793
borno | -.0103717 .0344786 -0.30 0.764 -.0779916 .0572483
adamawa | -.0393089 .0328901 -1.20 0.232 -.1038136 .0251957
gombe | -.0427407 .0284893 -1.50 0.134 -.0986144 .013133
bauchi | -.0357631 .0287795 -1.24 0.214 -.0922058 .0206797
kano | -.0149118 .0267639 -0.56 0.577 -.0674016 .0375781
kaduna | 0 (omitted)
kebbi | .0381979 .0296662 1.29 0.198 -.0199839 .0963798
niger | .0069699 .0320272 0.22 0.828 -.0558424 .0697822
abuja | .0504854 .0401723 1.26 0.209 -.0283013 .129272
nasarawa | .0686014 .0421795 1.63 0.104 -.0141218 .1513246
plateau | -.0435662 .0380323 -1.15 0.252 -.1181558 .0310233
taraba | .0768427 .0306792 2.50 0.012 .0166742 .1370113
benue | .0243015 .0385983 0.63 0.529 -.0513981 .1000012
kogi | .1786983 .0367717 4.86 0.000 .106581 .2508156
kwara | -.0002752 .0455985 -0.01 0.995 -.0897036 .0891533
oyo | .0816014 .0405837 2.01 0.044 .0020081 .1611948
osun | -.0841244 .0447581 -1.88 0.060 -.1719047 .0036558
ekiti | .0978379 .0656724 1.49 0.136 -.03096 .2266357
ondo | .0065026 .0508841 0.13 0.898 -.0932922 .1062973
edo | -.0097725 .055788 -0.18 0.861 -.1191849 .09964
anambra | .1023958 .0540376 1.89 0.058 -.0035836 .2083752
enugu | .0288383 .0501233 0.58 0.565 -.0694644 .1271411
ebonyi | -.0120409 .0507757 -0.24 0.813 -.1116231 .0875412
cross river | .1230477 .0564907 2.18 0.030 .0122572 .2338382
akwa ibom | .0265016 .0506741 0.52 0.601 -.0728814 .1258845
abia | -.004264 .0582497 -0.07 0.942 -.1185042 .1099763
imo | -.1218619 .0601408 -2.03 0.043 -.2398111 -.0039127
rivers | .00535 .0515217 0.10 0.917 -.0956954 .1063953
bayelsa | .174819 .0441338 3.96 0.000 .0882631 .2613748
delta | .0734782 .0554011 1.33 0.185 -.0351753 .1821317
lagos | -.0263495 .0557287 -0.47 0.636 -.1356456 .0829466
ogun | -.0620484 .0489006 -1.27 0.205 -.1579531 .0338563
Example with Sokoto as baseline state (not depicted) and Adamawa as omitted state - BH=1 if observation from Borno, Yobe, Adamawa state
Code:
surveystate |
zamfara | -.0461641 .0322722 -1.43 0.153 -.1094568 .0171286
katsina | .1004389 .0270741 3.71 0.000 .0473407 .1535372
jigawa | .0171015 .0300797 0.57 0.570 -.0418913 .0760942
yobe | -.0338448 .0318974 -1.06 0.289 -.0964025 .0287128
borno | .0282791 .0356248 0.79 0.427 -.0415887 .098147
adamawa | 0 (omitted)
gombe | .0653065 .0268831 2.43 0.015 .012583 .1180301
bauchi | .0723759 .0271817 2.66 0.008 .0190667 .1256852
kano | .0932255 .0256643 3.63 0.000 .0428923 .1435587
kaduna | .1099206 .0305516 3.60 0.000 .0500024 .1698388
kebbi | .0379399 .029656 1.28 0.201 -.0202219 .0961018
niger | .0069781 .0320587 0.22 0.828 -.055896 .0698523
abuja | .051153 .0401501 1.27 0.203 -.02759 .1298959
nasarawa | .0691163 .0421202 1.64 0.101 -.0134906 .1517232
plateau | -.0430699 .0380091 -1.13 0.257 -.1176141 .0314742
taraba | .0766932 .0306949 2.50 0.013 .0164938 .1368925
benue | .0248443 .0386022 0.64 0.520 -.0508629 .1005516
kogi | .1793773 .0367779 4.88 0.000 .107248 .2515067
kwara | .0000419 .0456494 0.00 0.999 -.0894865 .0895703
oyo | .082416 .0405685 2.03 0.042 .0028523 .1619797
osun | -.0832705 .0447555 -1.86 0.063 -.1710456 .0045047
ekiti | .0986303 .065534 1.51 0.132 -.0298962 .2271567
ondo | .0073753 .0508517 0.15 0.885 -.0923559 .1071065
edo | -.0086701 .0557107 -0.16 0.876 -.1179308 .1005906
anambra | .1031334 .0540501 1.91 0.057 -.0028706 .2091375
enugu | .029661 .0501349 0.59 0.554 -.0686645 .1279864
ebonyi | -.0117069 .0507315 -0.23 0.818 -.1112023 .0877885
cross river | .1240106 .0565273 2.19 0.028 .0131484 .2348728
akwa ibom | .0274592 .0506767 0.54 0.588 -.0719288 .1268473
abia | -.0029655 .0582343 -0.05 0.959 -.1171755 .1112445
imo | -.1205495 .0601398 -2.00 0.045 -.2384966 -.0026024
rivers | .0064126 .0515313 0.12 0.901 -.0946515 .1074766
bayelsa | .1760098 .0441195 3.99 0.000 .0894819 .2625376
delta | .0747945 .0553462 1.35 0.177 -.0337514 .1833404
lagos | -.0253275 .0556495 -0.46 0.649 -.1344683 .0838133
ogun | -.0611065 .0488985 -1.25 0.212 -.157007 .034794
Comment