回归时分别采取两种方法输入变量

1.

SER05 SER02 SER03 SER04

2.

SER05 C SER02 SER03 SER04

得到的结果大相径庭

如下：

第一种方法的结果

Dependent Variable: SER05
Method: Least Squares
Date: 02/13/07 Time: 20:01
Sample(adjusted): 1 422
Included observations: 378
Excluded observations: 44 after adjusting endpoints
Variable	Coefficient	Std. Error	t-Statistic	Prob.
SER02	7.320717	1.451472	5.043649	0
SER03	0.86629	0.175852	4.926244	0
SER04	8.437347	3.025258	2.788967	0.0056
R-squared	-4.55595	Mean dependent var		4.78809
Adjusted R-squared	-4.585582	S.D. dependent var		1.933206
S.E. of regression	4.568908	Akaike info criterion		5.88433
Sum squared resid	7828.095	Schwarz criterion		5.91556
Log likelihood	-1109.138	Durbin-Watson stat		0.957205

第二种方法的结果：

Dependent Variable: SER05
Method: Least Squares
Date: 02/13/07 Time: 21:50
Sample(adjusted): 1 422
Included observations: 378
Excluded observations: 44 after adjusting endpoints
Variable	Coefficient	Std. Error	t-Statistic	Prob.
C	5.449014	0.11629	46.85727	0
SER02	-1.631895	0.586481	-2.782518	0.0057
SER03	-0.583128	0.073958	-7.884591	0
SER04	-0.790018	1.172357	-0.673871	0.5008
R-squared	0.191344	Mean dependent var		4.78809
Adjusted R-squared	0.184857	S.D. dependent var		1.933206
S.E. of regression	1.7454	Akaike info criterion		3.96237
Sum squared resid	1139.362	Schwarz criterion		4.004009
Log likelihood	-744.888	F-statistic		29.49859
Durbin-Watson stat	1.996412	Prob(F-statistic)		0