i want to know how i can conduct a Hausman Test with Eviews concretely. I can read chinese, but can't write it. Thanks!
请注明：姓名-公司-职位
以便审核进群资格，未注明则拒绝
在eviews5.1中，先做固定效应，再做随机效应，然后进行hausman test.这个过程既可以通过下拉菜单点击完成，有可通过命令输入完成。
另外，用stata也很方便！
谢拉!今天在家,终于可以用中文了!再请问下,Hausman检验出来怎样判定呢?谢谢!
我的结果是:
Redundant Fixed Effects Tests
Pool: AA
Test cross-section fixed effects
Effects Test
Statistic
d.f.
Prob.
Cross-section F
15.470864
(306,1836)
0.0000
Cross-section fixed effects test equation:
Dependent Variable: ROA?
Method: Panel EGLS (Cross-section weights)
Date: 04/28/07 Time: 19:56
Sample: 1999 2005
Included observations: 7
Cross-sections included: 307
Total pool (balanced) observations: 2149
Use pre-specified GLS weights
Variable
Coefficient
Std. Error
t-Statistic
C
0.068396
0.006325
10.81384
SSP?
-0.037504
0.014492
-2.587954
0.0097
SSP2?
0.055023
0.015418
3.568668
0.0004
BW?
-0.003470
0.001070
-3.244181
0.0012
UMS?
0.000991
0.000803
1.234656
0.2171
W?
0.000487
0.000266
1.827696
0.0677
VG?
-0.006620
0.000461
-14.37483
Correlated Random Effects - Hausman Test
Test cross-section random effects
Test Summary
Chi-Sq. Statistic
Chi-Sq. d.f.
Cross-section random
76.696408
6
Cross-section random effects test comparisons:
Fixed
Random
Var(Diff.)
-0.103757
-0.094212
0.001973
0.8299
0.167572
0.143146
0.002959
0.6534
-0.043271
-0.015296
0.000015
0.022703
0.016645
0.000006
0.0105
0.000088
0.000149
0.000000
0.1722
-0.000561
-0.001044
Cross-section random effects test equation:
Method: Panel Least Squares
Date: 04/28/07 Time: 19:57
0.198540
0.029071
6.829446
0.060972
-1.701705
0.0890
0.070587
2.373977
0.0177
0.004977
-8.695024
0.003212
7.068805
8.75E-05
0.000154
0.568135
0.5700
0.000255
-2.197704
0.0281
Effects Specification
Cross-section fixed (dummy variables)
R-squared
0.348088
Mean dependent var
0.025812
Adjusted R-squared
0.237305
S.D. dependent var
0.065441
S.E. of regression
0.057151
Akaike info criterion
-2.752343
Sum squared resid
5.996879
Schwarz criterion
-1.926111
Log likelihood
3270.393
F-statistic
3.142082
Durbin-Watson stat
1.905307
Prob(F-statistic)
同问~~~期待高人解答~~
[此贴子已经被作者于2007-4-30 16:14:08编辑过]
看看Greene（5th）关于Hausman检验的hypothesis就知道怎么判断了。
Hausman检验显著用FEM，不显著时用REM。
扫码加好友，拉您进群
全部版块
我的主页
收藏
