Since I am not very familiar with the margin analysis, so it will be really appreciated if you can give me some ideas about the marginsplot.
I firstly run the fixed effect regression:
Code:
. xtreg coverage high i.high#i.year delta_child No_child_k_2,fe
note: 1.high#2015.year omitted because of collinearity
Fixed-effects (within) regression Number of obs = 236
Group variable: region2 Number of groups = 10
R-sq: Obs per group:
within = 0.9575 min = 20
between = 0.2169 avg = 23.6
overall = 0.7916 max = 24
F(25,201) = 181.02
corr(u_i, Xb) = -0.0885 Prob > F = 0.0000
------------------------------------------------------------------------------
coverage | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
high | 13.50749 .4410619 30.62 0.000 12.63779 14.37719
|
high#year |
0 2005 | 2.697003 .4492615 6.00 0.000 1.811132 3.582873
0 2006 | 4.109897 .4475569 9.18 0.000 3.227388 4.992406
0 2007 | 3.661136 .4424459 8.27 0.000 2.788705 4.533567
0 2008 | 3.483452 .4315488 8.07 0.000 2.632509 4.334396
0 2009 | 6.500022 .4366065 14.89 0.000 5.639105 7.360939
0 2010 | 7.397191 .4361316 16.96 0.000 6.537211 8.257171
0 2011 | 9.491619 .4361666 21.76 0.000 8.63157 10.35167
0 2012 | 10.58361 .4381868 24.15 0.000 9.71958 11.44765
0 2013 | 11.03185 .44718 24.67 0.000 10.15009 11.91362
0 2014 | 10.58706 .4626824 22.88 0.000 9.674727 11.4994
0 2015 | 10.29855 .4410619 23.35 0.000 9.428844 11.16825
1 2004 | -1.751561 .4299237 -4.07 0.000 -2.5993 -.9038216
1 2005 | -1.696726 .4299592 -3.95 0.000 -2.544536 -.848917
1 2006 | -2.01184 .427766 -4.70 0.000 -2.855324 -1.168355
1 2007 | -1.634555 .4257643 -3.84 0.000 -2.474093 -.7950178
1 2008 | -1.244545 .4264252 -2.92 0.004 -2.085386 -.4037044
1 2009 | -.0314732 .419265 -0.08 0.940 -.8581953 .7952488
1 2010 | .5597646 .419487 1.33 0.184 -.2673953 1.386924
1 2011 | .7301144 .4200139 1.74 0.084 -.0980844 1.558313
1 2012 | .9011659 .4191334 2.15 0.033 .0747035 1.727628
1 2013 | .7714839 .4194319 1.84 0.067 -.0555672 1.598535
1 2014 | .3475823 .4263088 0.82 0.416 -.493029 1.188193
1 2015 | 0 (omitted)
|
delta_child | .0204118 .018742 1.09 0.277 -.0165444 .0573679
No_child_k_2 | -.0021787 .004447 -0.49 0.625 -.0109475 .0065901
_cons | 82.65002 .6451935 128.10 0.000 81.3778 83.92223
-------------+----------------------------------------------------------------
sigma_u | 2.0873483
sigma_e | .93597554
rho | .83259355 (fraction of variance due to u_i)
------------------------------------------------------------------------------
F test that all u_i=0: F(9, 201) = 99.18 Prob > F = 0.0000Now I run:
Code:
. margins year#high
Predictive margins Number of obs = 236
Model VCE : Conventional
Expression : Linear prediction, predict()
------------------------------------------------------------------------------
| Delta-method
| Margin Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
year#high |
2004 0 | 82.36253 .3197892 257.55 0.000 81.73575 82.9893
2004 1 | 94.11846 .3041389 309.46 0.000 93.52235 94.71456
2005 0 | 85.05953 .3186661 266.92 0.000 84.43496 85.6841
2005 1 | 94.17329 .3053715 308.39 0.000 93.57477 94.77181
2006 0 | 86.47242 .3175103 272.35 0.000 85.85011 87.09473
2006 1 | 93.85818 .3026694 310.10 0.000 93.26496 94.4514
2007 0 | 86.02366 .3156681 272.51 0.000 85.40496 86.64236
2007 1 | 94.23546 .2996899 314.44 0.000 93.64808 94.82284
2008 0 | 85.84598 .3020403 284.22 0.000 85.25399 86.43797
2008 1 | 94.62547 .3020403 313.29 0.000 94.03348 95.21746
2009 0 | 88.86255 .2961302 300.08 0.000 88.28214 89.44295
2009 1 | 95.83854 .2961302 323.64 0.000 95.25814 96.41895
2010 0 | 89.75972 .2965945 302.63 0.000 89.1784 90.34103
2010 1 | 96.42978 .2965945 325.12 0.000 95.84847 97.0111
2011 0 | 91.85415 .2976081 308.64 0.000 91.27084 92.43745
2011 1 | 96.60013 .2976081 324.59 0.000 96.01683 97.18343
2012 0 | 92.94614 .2974581 312.47 0.000 92.36313 93.52915
2012 1 | 96.77118 .2974581 325.33 0.000 96.18818 97.35419
2013 0 | 93.39438 .3015762 309.69 0.000 92.8033 93.98546
2013 1 | 96.6415 .3015762 320.45 0.000 96.05042 97.23258
2014 0 | 92.94959 .3130885 296.88 0.000 92.33595 93.56323
2014 1 | 96.2176 .3130885 307.32 0.000 95.60396 96.83124
2015 0 | 92.66107 .2976214 311.34 0.000 92.07775 93.2444
2015 1 | 95.87002 .2976214 322.12 0.000 95.28669 96.45334
------------------------------------------------------------------------------1. No confidence interval of the two groups is intersected,
2. the CI is very narrowed. I have no clues about what this means and is that normal, or meaning my diagram has some problem？
Thank you so much, and very looking forward to your fantastic ideas! Array
0 Response to Marginsplot: two groups with 10 study years - no intersection of CI and very narrow 95% CIs
Post a Comment