求 sample size code

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

1:1 matched case-control study design,

how to calculate the sample size needed to guarteen alpha=0.05, beta=0.8 in SAS or STATA?

Thanks,

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0 回帖

关键词：Sample AMPL code size amp design needed

sdu0632 发表于 2010-5-24 01:06
1:1 matched case-control study design,

how to calculate the sample size needed to guarteen alpha=0.05, beta=0.8 in SAS or STATA?

Thanks,

It can be done in proc power/proc glmpower. You may refer to the documents for details.

Here is a simple example.

proc power;
      onesamplemeans
         mean   = 0.5
         ntotal = .
         stddev =  1
         power  = 0.9
         alpha  =0.05
        sides=2;
   run;

is there any special concern about matched case-control study?

personal quesiton: what do you do in USA? I am still a phd student, using SAS a lot.

sdu0632 发表于 2010-5-24 11:04
is there any special concern about matched case-control study?

personal quesiton: what do you do in USA? I am still a phd student, using SAS a lot.

You need a statistic. T test? ChiSq test? Case-control is about the data.

chisq test, 2 by 2 table, 1:1 matched case control study.

5# sdu0632

In proc power, using twosamplefreq or pairedfreq.
Just for reminding, considering the dependence between pairs, probably McNemar’s Test is more powerful for Matched Case-Control Studies because ChiSq test required independent observations. JingJu

thanks, Jinan