<P><STRONG><FONT face=Verdana color=#000000>按arlionn说的控制个体效果的操作，</FONT></STRONG></P>
<P>tsset province year</P>
<P>xtreg y x1-x7,fe</P>
<P>xtdata y x1-x7 year,fe clear</P>
<P>abar, lags(3) （发现存在一阶自相关）</P>
<P>xtgls y x1 x2 x3 x4 x5 x6 x7, panels(correlated) corr(ar1)Cross-sectional time-series FGLS regression</P>
<P>Coefficients:generalized least squares<BR>Panels: heteroskedastic with cross-sectional correlation<BR>Correlation: common AR(1) coefficient for all panels(0.0441)</P>
<P>Estimated covariances = 66 Number of obs = 99<BR>Estimated autocorrelations = 1 Number of groups = 11<BR>Estimated coefficients = 8 Time periods = 9<BR> Wald chi2(7) =1.81e+07<BR> Prob > chi2 = 0.0000</P>
<P>------------------------------------------------------------------------------<BR> y | Coef. Std. Err. z P>|z| [95% Conf. Interval]<BR>-------------+----------------------------------------------------------------<BR> x1 | 3.407846 .0361053 94.39 0.000 3.337081 3.478611<BR> x2 |-.0825687 .0001589-519.69 0.000 -.0828801 -.0822573<BR> x3 | .1726899 .000246 702.00 0.000 .1722077 .173172<BR> x4 | .1446033 .0002017 716.93 0.000 .144208 .1449986<BR> x5 | .244987 .000687 356.62 0.000 .2436406 .2463334<BR> x6 | 1.392012 .0089145 156.15 0.000 1.37454 1.409485<BR> x7 |-.1784017 .0006139-290.61 0.000 -.1796049 -.1771985<BR> _cons |-30.99539 .3388195 -91.48 0.000 -31.65947 -30.33132<BR>------------------------------------------------------------------------------<BR>为什么z检验值异常的大呢？<STRONG><FONT face=Verdana>arlionn能帮忙解释一下这个结果是否正确么？其他达人如知道也请指点！多谢了！</FONT></STRONG></P>