您好!我不太会用Eviews进行异方差检验,我用的怀特检验法得到了如下图的结果,不太懂到底有没有异方差White Heteroskedasticity Test:F-statistic2.791081 Probability0.075217Obs*R-s

您好!我不太会用Eviews进行异方差检验,我用的怀特检验法得到了如下图的结果,不太懂到底有没有异方差
White Heteroskedasticity Test:

F-statistic2.791081 Probability0.075217
Obs*R-squared8.193354 Probability0.084747


Test Equation:
Dependent Variable: RESID^2
Method: Least Squares
Date: 01/05/12 Time: 16:00
Sample: 1994 2010
Included observations: 17

VariableCoefficient Std. Err t-Statistic Prob.

C478.2814 20146.31 0.023740 0.9814
X1-0.2916456.359985 -0.0458560.9642
X1^20.000285 0.000251 1.1345790.2787
X2-18.7784172.22224 -0.2600090.7993
X2^2-0.0206940.024121 -0.8579550.4077

R-squared 0.481962 Mean dependent var7621.512
Adjusted R-squared0.309283 S.D. dependent var11733.10
S.E. of regression9751.312 Akaike info criterion21.44812
Sum squared resid1.14E+09 Schwarz criterion 21.69318
Log likelihood-177.3090 F-statistic2.791081
Durbin-Watson stat2.199710 Prob(F-statistic)0.075217

我也刚好在做这个噢,用N*R^2得出卡方值,在你这里就是17* 0.481962=8.193354,然后查表,你这里是自由度为4,如果在显著性水平为0.05的情况下卡方临界值为9.4877,临界值大于你的卡方值,所以应该是认为不存在异方差的.

修正之后更加不对。 没法检验现实不存在异方差现象了。。。 显著水平是多少？ 如果α是0.05的话 P值<α ，拒绝原假设，存在r阶ARCH效应