Heteroskedasticity Test: White   
   
F-statistic 2.910883     Prob. F(5,25)  0.0332
Obs*R-squared 11.40674     Prob. Chi-Square(5)  0.0439
Scaled explained SS 3.074024     Prob. Chi-Square(5)  0.6886
   
   
Test Equation:   
Dependent Variable: RESID^2   
Method: Least Squares   
Date: 12/25/15   Time: 10:05   
Sample: 1984 2014   
Included observations: 31   
   
Variable Coefficient Std. Error t-Statistic Prob.  
   
C -2.36E+08 2.06E+09 -0.114603 0.9097
X5 -226316.5 138018.1 -1.639759 0.1136
X5^2 2.487956 1.056798 2.354239 0.0267
X5*X3 1928.765 1288.550 1.496848 0.1470
X3 9767784. 36865864 0.264955 0.7932
X3^2 -63482.46 164248.9 -0.386501 0.7024
   
R-squared 0.367959     Mean dependent var  43178402
Adjusted R-squared 0.241551     S.D. dependent var  35676179
S.E. of regression 31070016     Akaike info criterion  37.51337
Sum squared resid 2.41E+16     Schwarz criterion  37.79092
Log likelihood -575.4572     Hannan-Quinn criter.  37.60384
F-statistic 2.910883     Durbin-Watson stat  1.035576
Prob(F-statistic) 0.033221   
   
卡方查表等于11.07   可不可以说不具有异方差
如果不能说明怎么修正，加权权重1/x3，1/x5,1/x3^2,1/x5^21/e,1/abs(e),1/e^2都尝试了结果反而更严重，可以取对数吗，
社会消费品零售总额Y、 商品零售价格指数X3、农村居民家庭平均每人纯收入X5、
Dependent Variable: LNY   
Method: Least Squares   
Date: 12/20/15   Time: 15:00   
Sample: 1984 2014   
Included observations: 31   
   
Variable Coefficient Std. Error t-Statistic Prob.  
   
C -2.071800 1.361718 -1.521460 0.1394
LNX3 0.557184 0.279390 1.994291 0.0559
LNX5 1.309577 0.016594 78.92085 0.0000
   
R-squared 0.996407     Mean dependent var  10.40254
Adjusted R-squared 0.996151     S.D. dependent var  1.290098
S.E. of regression 0.080040     Akaike info criterion  -2.120812
Sum squared resid 0.179380     Schwarz criterion  -1.982039
Log likelihood 35.87259     Hannan-Quinn criter.  -2.075576
F-statistic 3882.919     Durbin-Watson stat  0.431313
Prob(F-statistic) 0.000000   
   
取对数结果合理吗