我用 glogit intrd id protect conglom inthamp ecohamp othhamp markinf publinf confinf size命令得出如下结论:
Weighted LS logistic regression for grouped data
Source | SS df MS Number of obs = 4397
-------------+------------------------------ F( 9, 4387) = 8.52
Model | 24.5512966 9 2.72792184 Prob > F = 0.0000
Residual | 1404.62942 4387 .320179945 R-squared = 0.0172
-------------+------------------------------ Adj R-squared = 0.0152
Total | 1429.18072 4396 .325109353 Root MSE = .56584
------------------------------------------------------------------------------
intrd | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
protect | -.0586423 .0399373 -1.47 0.142 -.1369396 .0196549
conglom | -.0455758 .0321546 -1.42 0.156 -.108615 .0174634
inthamp | -.0601496 .0393465 -1.53 0.126 -.1372886 .0169893
ecohamp | .0525813 .0291548 1.80 0.071 -.0045768 .1097395
othhamp | .0117128 .0243618 0.48 0.631 -.0360487 .0594743
markinf | .0127769 .0037188 3.44 0.001 .0054862 .0200675
publinf | -.0158026 .0040372 -3.91 0.000 -.0237175 -.0078877
confinf | .0070575 .0041912 1.68 0.092 -.0011594 .0152745
size | 1.775859 .2637298 6.73 0.000 1.258816 2.292903
_cons | -8.348138 .0461562 -180.87 0.000 -8.438627 -8.257648
------------------------------------------------------------------------------
几个问题:
1、我的样本数是 obs: 7,570 为什么得出结论中 Number of obs = 4397?
2、在做logit模型前要不要检验变量之间的多重共线性、异方差性等,要的话命令是什么?对不起我是一上来就学logit的,对回归做的不是很多
3、在做出来的模型后是不是主要p值全为0,或是大多数接近于0就可以通过模型的检验了(当然还有t检验量要大于一个数,这个数是多少,有没有一个标准)?
4、我尝试着用noconstant命令去掉常数项,但是不行,不知能不能去掉,要是能的话,怎么去?
谢谢各位高手指点了!祝端午节快乐!
附:su命令结果如下:
Variable | Obs Mean Std. Dev. Min Max
-------------+--------------------------------------------------------
id | 7570 3954.335 2264.061 1 7897
code | 7570 29.96314 8.503422 13 42
extrd | 4398 1.704411 .4563589 1 2
intrd | 4398 1.096407 .2951828 1 2
rd | 7570 .5384412 .498553 0 1
-------------+--------------------------------------------------------
techno | 7570 .1981506 .3986326 0 1
machi | 7570 .4101717 .4918972 0 1
technology | 7570 .447424 .4972609 0 1
trial | 7570 .2376486 .4256709 0 1
market | 7570 .3819022 .4858849 0 1
-------------+--------------------------------------------------------
train | 7570 .4639366 .4987307 0 1
downdream | 7570 .4970938 .5000246 0 1
protect | 7570 .5858653 .4926045 0 1
conglom | 7570 .1792602 .3835954 0 1
inthamp | 7570 .2412153 .8929741 0 4
-------------+--------------------------------------------------------
ecohamp | 7570 .2529723 .931145 0 4
othhamp | 7570 1.502246 1.487272 0 7
markinf | 7570 5.903831 5.506143 0 20
publinf | 7570 5.82959 5.573115 0 16
confinf | 7570 5.687318 5.409129 0 16
-------------+--------------------------------------------------------
size | 7570 .0068146 .0222115 8.41e-06 .9044183
.