∑e2=6030148 ∑e2=2495840 求F统计量为 =4.1390 ∑26030148 给定a=0.05,查F分布表,得临界值为F05(20,20)=212 c比较临界值与F统计量值,有F=41390>F0(20.,20)=212,说明该模型的随机误 差项存在异方差 其次,用 White法进行检验。具体结果见下表 White Heteroskedasticity Test: Fstatisti 6.301373 0.003370 Obs*R-squared 10.86401 Probability 0.004374 Test Equation Dependent Variable: RESID2 Method: Least Squares Date:08/05/05Tme:12:37 Sample: 1 60 included observations: 60 Coefficient std error t-statistic Prob C 0.1659771.6198560.1024640.918 X^2 0.0018000.0045870.3924690.6962 0. 181067 Mean dependent var 78.86225 Adjusted R-squared 0. 152332 S.D. dependent var 111.1375 S.E. of regressio Akaike info criterion 12.14285 Sum squared resid 596790.5 Schwarz criterion 12.24757 Log likelihood 361.2856 F-statistic 6.301373 Durbin-Watson stat 0.937366 Prob(F-statistic) 0.003370 给定a=005,在自由度为2下查卡方分布表,得x2=59915 比较临界值与卡方统计量值,即nR2=10.8640>x2=5915,同样说明模型中的随机误 差项存在异方差。 (2)用权数W1厂’作加权最小二乘估计,得如下结果 Dependent vanable:Y8 2 1 2 2 603.0148 2495.840 e e = =   求 F 统计量为 2 2 2 1 2495.84 4.1390 603.0148 e F e = = =   给定  = 0.05 ，查 F 分布表，得临界值为 0.05 F (20,20) 2.12 = 。 c.比较临界值与 F 统计量值，有 F =4.1390> 0.05 F (20,20) 2.12 = ，说明该模型的随机误 差项存在异方差。 其次，用 White 法进行检验。具体结果见下表 White Heteroskedasticity Test: F-statistic 6.301373 Probability 0.003370 Obs*R-squared 10.86401 Probability 0.004374 Test Equation: Dependent Variable: RESID^2 Method: Least Squares Date: 08/05/05 Time: 12:37 Sample: 1 60 Included observations: 60 Variable Coefficient Std. Error t-Statistic Prob. C -10.03614 131.1424 -0.076529 0.9393 X 0.165977 1.619856 0.102464 0.9187 X^2 0.001800 0.004587 0.392469 0.6962 R-squared 0.181067 Mean dependent var 78.86225 Adjusted R-squared 0.152332 S.D. dependent var 111.1375 S.E. of regression 102.3231 Akaike info criterion 12.14285 Sum squared resid 596790.5 Schwarz criterion 12.24757 Log likelihood -361.2856 F-statistic 6.301373 Durbin-Watson stat 0.937366 Prob(F-statistic) 0.003370 给定  = 0.05 ，在自由度为 2 下查卡方分布表，得 2  = 5.9915 。 比较临界值与卡方统计量值，即 2 2 nR =  = 10.8640 5.9915  ，同样说明模型中的随机误 差项存在异方差。 （2）用权数 1 W1 X = ，作加权最小二乘估计，得如下结果 Dependent Variable: Y