计量操作~最终版(夏璐明).docVIP

注：操作结果有修改，已用红字注明 1(1)OLS操作结果为 Dependent Variable: Y Method: Least Squares Date: 12/15/11 Time: 17:20 Sample: 1 20 Included observations: 20 Variable Coefficient Std. Error t-Statistic Prob. C 379.7981 80.73423 4.704302 0.0002 X 0.875215 0.124766 7.014854 0.0000 R-squared 0.732175 Mean dependent var 937.3650 Adjusted R-squared 0.717296 S.D. dependent var 119.0540 S.E. of regression 63.30086 Akaike info criterion 11.22831 Sum squared resid 72125.98 Schwarz criterion 11.32789 Log likelihood -110.2831 F-statistic 49.20818 Durbin-Watson stat 2.023366 Prob(F-statistic) 0.000002 用普通最小二乘法估计家庭的消费函数为Y=379.7981+0.875215X (2)检验序列相关性：DW检验 由回归结果DW=2.023366，查DW分布表知，在5%的显著性水平下，样本为20，解释变量数为1，DW统计量的下界dl=1.201,上界du=1.411。du=1.411DW=2.0233664-du=2.589，故原回归模型不存在随机干扰项的一阶自相关性。 （3）检验异方差性 图示检验法 模型OLS回归得到的残差平方项E2与X的散点图表明可能存在单调递增型异方差。 G-Q检验 将原始数据X排成升序，去掉中间4个数据，得到两个容量为8的子样本。对两个子样本分别做OLS回归，得到各自的残差平方和RSS1和RSS2。 子样本1的回归结果为 Dependent Variable: Y Method: Least Squares Date: 12/15/11 Time: 18:15 Sample: 1 8 Included observations: 8 Variable Coefficient Std. Error t-Statistic Prob. C 369.5124 135.0129 2.736868 0.0339 X 0.903798 0.253083 3.571145 0.0118 R-squared 0.680053 Mean dependent var 849.0625 Adjusted R-squared 0.626728 S.D. dependent var 64.83055 S.E. of regression 39.60887 Akaike info criterion 10.40830 Sum squared resid 9413.174 Schwarz criterion 10.42816 Log likelihood -39.63320 F-statistic 12.75308 Durbin-Watson stat 1.997218 Prob(F-statistic) 0.011768 Y=369.5124 + 0.903798X (2.736868) ( 3.571145) R2=0.680053 RSS1=9413.174 子样本2的回归结果为 Dependent Variable: Y Method: Least Squares Date: 12/15/11 Time: 18:24 Sample: 13 20 Included observations: 8 Variable Coefficient Std. Error t-Statistic Prob. C 211.7592 354.4028 0.597510 0.5720 X 1.090613 0.471408 2.313525 0.0600 R-squared 0.471477 Mean dependent var 1027.812 Adjusted R-squared 0.383390 S.D. dependent var 123.8269 S.E. of regres