SAS编程计算率的95%置信区间。急等答案，谢谢！

shiyiming

请问一下如何用SAS编程计算某个率的95%置信区间？例如有100个受访者，对A,B两种物品做出好，一般，差的评价。100人中队两种物品的评价相同的比例为一致率，请问如何用SAS来计算这个一致率的95%置信区间？谢谢，急等答案！

shiyiming

[code:1xe4m806]


data ads;
do a=1 to 3;
    do b=1 to 3;
        input w@@;
        output;
   end;
end;
cards;
48 24 8 10 25 25 10 25 25
;

proc freq;
  weight w;
  tables a*b/agree;
  output out=bds kappa;
run;
data cds;
  set bds;
  u=_kappa_/E_kappa;
  p=2*(1-probnorm(u));
  u=round(u,0.001);
  p=round(p,0.000001);
  proc print;
run;

[/code:1xe4m806]

shiyiming

[quote="MerlinZHOU":ngdzlwp1][code:ngdzlwp1]


data ads;
do a=1 to 3;
    do b=1 to 3;
        input w@@;
        output;
   end;
end;
cards;
48 24 8 10 25 25 10 25 25
;

proc freq;
  weight w;
  tables a*b/agree;
  output out=bds kappa;
run;
data cds;
  set bds;
  u=_kappa_/E_kappa;
  p=2*(1-probnorm(u));
  u=round(u,0.001);
  p=round(p,0.000001);
  proc print;
run;

[/code:ngdzlwp1][/quote:ngdzlwp1]

100人评价，这回复里这data步的意思是~？？求大牛们解释~？

shiyiming

[quote="MerlinZHOU":2ux7th9z][code:2ux7th9z]


data ads;
do a=1 to 3;
    do b=1 to 3;
        input w@@;
        output;
   end;
end;
cards;
48 24 8 10 25 25 10 25 25
;

proc freq;
  weight w;
  tables a*b/agree;
  output out=bds kappa;
run;
data cds;
  set bds;
  u=_kappa_/E_kappa;
  p=2*(1-probnorm(u));
  u=round(u,0.001);
  p=round(p,0.000001);
  proc print;
run;

[/code:2ux7th9z][/quote:2ux7th9z]

100人评价，这回复里这data步的意思是~？？求大牛们解释~？

shiyiming

原始数据入下
[code:1ya356sp]
                 b
a   
       1      2      3
1    48    24      8
2    10    25      25
3    10    25      25


[/code:1ya356sp]
通俗的解释是w的观测值被分成a,b两类
采用上述方式数据输入是为了评价a和b 之间的关系或a b对w的影响等。