用stata做双向固定效应时,R平方数值特别小,算正常现象吗?
第一次做双向固定效应,然后看到那么小的R^2,感觉都要哭了。
各位大神,帮忙看看这样的R^2是正常现象吗?
(蓝色字体部分)还是只看回归的F统计量和t值,R^2直接忽略?
还有就是中间2003年数据omitted是什么意思?
谢谢啦
. xi: xtreg y x1 x2 x3 i.year ,fe i(code)
i.year _Iyear_2000-2007 (naturally coded; _Iyear_2000 omitted)
note: _Iyear_2003 omitted because of collinearity
Fixed-effects (within) regression Number of obs = 8406
Group variable: code Number of groups = 2407
R-sq: within = 0.0729 Obs per group: min = 1
between = 0.0012 avg = 3.5
overall = 0.0089 max = 8
F(9,5990) = 52.34
corr(u_i, Xb) = -0.0609 Prob > F = 0.0000
------------------------------------------------------------------------------
y | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
x1 | .0494374 .0076836 6.43 0.000 .0343749 .0645
x2 | .0051569 .0044847 1.15 0.250 -.0036348 .0139486
x3 | .0204274 .0494387 0.41 0.679 -.0764903 .117345
_Iyear_2001 | .009736 .0479477 0.20 0.839 -.0842589 .1037308
_Iyear_2002 | .2615747 .0487299 5.37 0.000 .1660465 .3571028
_Iyear_2003 |
0 (omitted)
_Iyear_2004 | .1924398 .0350422 5.49 0.000 .1237444 .2611352
_Iyear_2005 | .3128351 .0422391 7.41 0.000 .2300313 .3956389
_Iyear_2006 | .3648249 .0364031 10.02 0.000 .2934617 .436188
_Iyear_2007 | .3814248 .0366246 10.41 0.000 .3096273 .4532223
_cons | 6.816618 .0507309 134.37 0.000 6.717167 6.916069
-------------+----------------------------------------------------------------
sigma_u | 1.2059948
sigma_e | .64069603
rho | .77988732 (fraction of variance due to u_i)
------------------------------------------------------------------------------
F test that all u_i=0: F(2406, 5990) = 11.33 Prob > F = 0.000