1 SPSS 作业7:曲线估计 为研究居民家庭教育支出和消费性支出之间的关系,收集到1990 年到2002 年全国人均消费性支出和教育支出的数据。首先绘制教育支出和消费性支出的散点图,如下所示: (一)教育支出的相关因素分析 曲线估计的基本操作: (1)选择菜单Analyze-Regression―Curve Estimation; (2)选择被解释变量到Dependent 框中,再选择菜单Graphs―Scatter;分别作简单散点图,矩阵散点图,结果如下: Graph 2 年人均消费性支出1000080006000400020000教育支出120010008006004002000 教育支出和年人均消费性支出的散点图 分析:观察散点图发现两变量之间呈非线性关系,其中,教育支出为被解释变量,消费性支出为解释变量,分析结果如下: 教育支出的曲线估计结果 Independent: X2 Dependent Mth Rsq d.f. F Sig.f b0 b1 b2 b3 3 X5 QUA 0.987 10 382.64 .000 252.698 -.1475 2.5E-05 - X5 CUB 0.994 9 516.46 .000 -41.314 .0754 -2.E-05 2.6E-09 X5 COM 0.995 11 2086.35 .000 20.9550 1.0004 - - X5 POW 0.954 11 229.58 .000 3.6E-05 1.8460 - - MODEL: MOD_2. Dependent variable.. X5 Method.. QUADRATI Listwise Deletion of Missing Data Multiple R .99353 R Square .98710 Adjusted R Square .98452 Standard Error 45.70690 Analysis of Variance: DF Sum of Squares Mean Square Regression 2 1598766.0 799383.00 Residuals 10 20891.2 2089.12 F = 382.64096 Signif F = .0000 -------------------- Variables in the Equation -------------------- Variable B SE B Beta T Sig T X2 -.147527 .025041 -1.134958 -5.892 .0002 X2**2 2.46018091E-05 2.2722E-06 2.085797 10.827 .0000 (Constant) 252.697890 57.792248 4.373 .0014 4 Dependent variable.. X5 Method.. CUBIC Listwise Deletion of Missing Data Multiple R .99711 R Square .99422 Adjusted R Square .99230 Standard Error 32.23848 Analysis of Variance: DF Sum of Squares Mean Square Regression 3 1610303.3 536767.78 Residuals 9 9353.9 1039.32 F = 516.46087 Signif F = .0000 -------------------- Variables in the Eq...