gen gapage=age-l.age
drop if gapage~=1 & gapage~=. & year~=2003
gen gapsch=sch-l.sch
drop if gapsch~=0 & gapsch~=. & year~=2003
gen gapgender=gender-l.gender
drop if gapgender~=0 & gapgender~=. & year~=2003
gen gapssch=ssch-l.ssch
drop if gapssch~=0 & gapssch~=.
gen gapsage=sage-l.sage
drop if gapsage~=1 & gapsage~=. & year~=2003
数据清理到这一步我感觉差不多了,于是做了一个固定效应模型。
xi: xtreg lwpro lwarea lwcferp lwothkp lwl i.gender if wl<1000, fe
其他的系数尚显合理,
------------------------------------------------------------------------------
lwpro | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
lwarea | .******* .******* ******* 0.000 .6734433 .7021303
lwcferp | .******* .******* ******* 0.000 .0580465 .0775423
lwothkp | .******* .******* ******* 0.000 .0333327 .0444901
lwl | .******* .******* ******* 0.000 .0547959 .0761264
_Igender_2 | -.0093158 .037758 -0.25 0.805 -.0833248 .0646933
sch | -.0071488 .0046509 -1.54 0.124 -.0162649 .0019674
_cons | 5.686381 .0432774 131.39 0.000 5.601554 5.771209
-------------+----------------------------------------------------------------
但是您看后两行,怎么还会有系数?我反复清理,总是做不下去,您看是否是我的清理思路不对?