Correlated Random Effects-Hausman Test

Pool:P1

Test cross-section random effects

Test Summary                  Chi-Sq. Statistic     Chi-Sq.df        Prob

Cross-section random        6.672827                        7            0.4637

Cross-section effects test comparisons:

Variable             Fixed             Random            Var(Diff.)            Prob.

RDS                 1.089123         1.864920           0.285049       0.1462

RDS(-1)           0.215631         0.666579           0.118302       0.1898

RDS(-2)           -0.046414        0.570829           0.245156       0.2125

ROA                0.420463          0.558179           0.003475      0.0195

TECH              0.020144         -0.011309          0.000820       0.2720

CAPEX             -0.168878       -0.168699         0.000242       0.9908

SIZE                -0.051592         -0.049723         0.000025     0.7111

Cross-section  random effects test equation

Dependent Vaviable:P1

Method :Panel  Least Squares

Date :08/18/08     Time:21:58

Sample(adjusted): 2005 2007

Included observations: 3 after adjustments

Cross-Sections included:66

TOtal pool (balanced) observations:198

                      Coefficient            Std.Error          t-Statistic          Prob

C                   1.348246             0.208303         6.472540        0.0000

RDS              1.089123             0.701883         1.551716       0.1233

RDS(-1)        0.215631             0.633167         0.340559       0.7340

RDS(-2)       -0.046414           0.705612         -0.065778        0.9477

ROA           0.420463             0.157262           2.673640       0.0085

TECH           0.020144              0.058867        0.342192       0.7328

CAPEX          -0.168878          0.051990           -3.248254      0.0015

SIZE            -0.051592           0.009770            -5.280929      0.0000

Effects Specification

R-squared       0.882230                Mean  dependent var 0.293530

Adjusted R-squared  0.814935       S.D dependent var   0.157350

S.E of regression   0.067789           Akaike  info criterion  -2.267404

Sum squared resid    0.574424      Schwarz  criterion   -1.055063

Log likelihood        297.4730           Hannan-Watson  stat   1.776688

F-statistic              13.00548            Dubin-Watson  stat      2.367504

Prob(F-statistic)    0.000000

这样的结果应采用固定效应模型还是随机效应模型？因初次用面板数据模型，所以不这么懂，帮忙分析一下，尽量详细一点，先谢谢了。

[此贴子已经被作者于2008-8-18 15:16:26编辑过]