<P>  做变截距固定系数面板模型的时候，用eviews5.0不知道该如何确定固定效应和随机效应。李子奈老师的书中说用SAS的Hausman检验可以做，可惜没有安装该软件，而且现在时间也比较紧。拜求！！我用eviews5.0做的估计结果如下，不知道这里Effects Specification   是什么意思。</P>
<P>Dependent Variable: JZ?    <BR>Method: Pooled EGLS (Cross-section random effects)    <BR>Date: 04/17/07   Time: 13:03    <BR>Sample: 1 7    <BR>Included observations: 7    <BR>Cross-sections included: 6    <BR>Total pool (balanced) observations: 42    <BR>Swamy and Arora estimator of component variances    <BR>    <BR>Variable Coefficient Std. Error t-Statistic Prob.  <BR>    <BR>C 280.6238 85.48916 3.282566 0.0021<BR>KZSR? 0.055957 0.007122 7.857278 0.0000<BR>Random Effects (Cross)    <BR>_00--C -28.51622   <BR>_01--C -82.13265   <BR>_02--C 100.3817   <BR>_03--C 16.12033   <BR>_04--C 59.55197   <BR>_05--C -65.40514   <BR>    <BR> Effects Specification   <BR>    <BR>Cross-section random S.D. / Rho   108.8416 0.1448<BR>Idiosyncratic random S.D. / Rho   264.5013 0.8552<BR>    <BR> Weighted Statistics   <BR>    <BR>R-squared 0.612836     Mean dependent var  511.7490<BR>Adjusted R-squared 0.603157     S.D. dependent var  414.6063<BR>S.E. of regression 261.1833     Sum squared resid  2728669.<BR>F-statistic 63.31533     Durbin-Watson stat  1.074382<BR>Prob(F-statistic) 0.000000   <BR>    <BR> Unweighted Statistics   <BR>    <BR>R-squared 0.590469     Mean dependent var  756.5081<BR>Sum squared resid 3060597.     Durbin-Watson stat  0.957864<BR>    <BR></P>
<P>          </P>