用EViews作横截面回归(711个样本),OLS结果见下图,c和要研究的解释变量x不显著
| | | | |
| | | | | |
| Variable | Coefficient | Std. Error | t-Statistic | Prob. |
| | | | | |
| | | | | |
| C | 0.016647 | 0.036383 | 0.457542 | 0.6474 |
| X | 0.002293 | 0.003776 | 0.607424 | 0.5438 |
| X1 | 0.004121 | 0.001893 | 2.176619 | 0.0298 |
| X2 | -0.000232 | 4.22E-05 | -5.484994 | 0.0000 |
| X3 | -0.000626 | 0.000132 | -4.753955 | 0.0000 |
| X4 | -0.682935 | 0.022803 | -29.94900 | 0.0000 |
| X5 | -8.10E-05 | 0.000117 | -0.689474 | 0.4908 |
| | | | | |
| | | | | |
| R-squared | 0.583459 | Mean dependent var | 0.046108 |
| Adjusted R-squared | 0.579909 | S.D. dependent var | 0.073209 |
| S.E. of regression | 0.047450 | Akaike info criterion | -3.248477 |
| Sum squared resid | 1.585069 | Schwarz criterion | -3.203517 |
| Log likelihood | 1161.833 | Hannan-Quinn criter. | -3.231110 |
| F-statistic | 164.3518 | Durbin-Watson stat | 1.927309 |
| Prob(F-statistic) | 0.000000 | | | |
| | | | | |
加权最小二乘后WLS,就t值就变得非常显著,拟合优度也高的离谱,是不是有问题啊,应从那几个方面考虑问题出在哪呢?
| Variable | Coefficient | Std. Error | t-Statistic | Prob. |
| | | | | |
| | | | | |
| C | 0.010903 | 0.001045 | 10.42947 | 0.0000 |
| X | 0.004323 | 7.72E-05 | 55.97742 | 0.0000 |
| X1 | 0.004934 | 5.27E-05 | 93.57255 | 0.0000 |
| X2 | -0.000372 | 1.61E-06 | -231.5402 | 0.0000 |
| X3 | -0.000962 | 1.72E-06 | -559.4482 | 0.0000 |
| X4 | -0.739531 | 0.000217 | -3404.976 | 0.0000 |
| X5 | 1.10E-05 | 1.94E-06 | 5.709232 | 0.0000 |
| | | | | |
| | | | | |
| | Weighted Statistics | | |
| | | | | |
| | | | | |
| R-squared | 0.999979 | Mean dependent var | 0.028508 |
| Adjusted R-squared | 0.999979 | S.D. dependent var | 0.298152 |
| S.E. of regression | 0.000968 | Akaike info criterion | -11.03316 |
| Sum squared resid | 0.000659 | Schwarz criterion | -10.98820 |
| Log likelihood | 3929.288 | Hannan-Quinn criter. | -11.01579 |
| F-statistic | 5571510. | Durbin-Watson stat | 2.081187 |
| Prob(F-statistic) | 0.000000 | | | |
| | | | | |
| | | | | |
| | Unweighted Statistics | | |
| | | | | |
| | | | | |
| R-squared | 0.566735 | Mean dependent var | 0.046108 |
| Adjusted R-squared | 0.563043 | S.D. dependent var | 0.073209 |
| S.E. of regression | 0.048393 | Sum squared resid | 1.648709 |
| Durbin-Watson stat | 1.908620 | | | |