请教SAS如何对“斜率=1”进行T检验，还有wald检验

是 否 +2 论坛币

k人 参与回答

经管之家送您一份

应届毕业生专属福利!

求职就业群

赵安豆老师微信：zhaoandou666

经管之家联合CDA

送您一个全额奖学金名额~ !

立即领取

感谢您参与论坛问题回答

经管之家送您两个论坛币！

+2 论坛币

请教SAS如何对“斜率=1”进行T检验，还有如何对“截距=0，斜率=1”进行WALD检验，都是一元回归，求高手指点

扫码加我 拉你入群

请注明：姓名-公司-职位

以便审核进群资格，未注明则拒绝

分享0 收藏1 回帖

关键词：wald检验 Wald t检验 高手指点 求高手 如何

1. data toluca;
   
2.         input LotSize WorkHrs @@;
   
3.         WorkHrs3 = WorkHrs - 3*LotSize;
   
4. cards;
   
5. 80 399 30 121 50 221 90 376 70 361 60 224 120 546 80 352 100 353 50 157 40 160 70 252
   
6. 90 389 20 113 110 435 100 420 30 212 50 268 90 377 110 421 30 273 90 468 40 244 80 342 70 323
   
7. ;
   
8. run;
   
9. proc reg data=toluca;
   
10.         model workhrs=lotsize;
    
11.         Test1: test lotsize=3;
    
12.         Test2: test Intercept=0, Lotsize=3;
    
13. run;
    
14. proc reg data=toluca;
    
15.         model WorkHrs3 = LotSize;
    
16. run;
    
17. proc reg data=toluca;
    
18.         model WorkHrs3 = /noint;
    
19. run;

复制代码

Here's an example.
In model 1, Test1 tests 'slope=3', the square root of the F-test statistic is the t-test statistics you want; Test2 tests 'Intercept=0 and Slope=3' and it's a Wald test.

Alternatively, subtract 3*LotSize from the response variable then run Model 2, you'll see the t-test in the 'Parameter Estimates' section of output.
Also you can compute the Wald test statistic manually as
(322793 - 54825)/2 / (54825/23),
where 322793=error sum of squares of model 3 (null model), 54825=error sum of squares from model 2 (full model), 2 = # of parameters in test, 23 = degrees of freedom in full model.