PPS抽样调查的SAS如何实现

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

现需要做一项调查，要求按照变量GDP进行PPS抽样，要求：按照GDP大小成比例抽取20个地区样本，请问采用SAS如何实现？谢谢
Region GDP
Beijing 11865.9
Tianjin 7500.8
Hebei 17026.6
Liaoning 15065.6
Shanghai 14900.9
Jiangsu 34061.2
Zhejiang 22832.4
Fujian 11949.5
Shandong 33805.3
Guangdong 39081.6
Hainan 1646.6
Shanxi 7365.7
Jilin 7203.2
Heilongjiang 8288.0
Anhui 10052.9
Jiangxi 7589.2
Henan 19367.3
Hubei 12831.5
Hunan 12930.7
Chongqing 6528.7
Sichuan 14151.3
Guizhou 3893.5
Yunnan 6168.2
Tibet 441.4
Shaanxi 8186.7
Gansu 3382.4
Qinghai 1081.3
Ningxia 1334.6
Xinjiang 4273.6
InnerMongolia 9725.8
Guangxi 7700.4

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5 回帖

关键词：抽样调查 如何实现 pps Chongqing sichuan 如何 样本

弱弱的问一下，PPS是什么意思？

surveyselect过程 1# wcguo94

leedx 发表于 2011-7-8 08:37
弱弱的问一下，PPS是什么意思？

PPS抽样调查法( Probability Proportionate to Size Sampling)又称，按规模大小成比例的概率抽样/PPS抽样

wcguo94 发表于 2011-7-7 10:45
现需要做一项调查，要求按照变量GDP进行PPS抽样，要求：按照GDP大小成比例抽取20个地区样本，请问采用SAS如何实现？谢谢
Region GDP
Beijing 11865.9
Tianjin 7500.8
Hebei 17026.6
Liaoning 15065.6
Shanghai 14900.9
Jiangsu 34061.2
Zhejiang 22832.4
Fujian 11949.5
Shandong 33805.3
Guangdong 39081.6
Hainan 1646.6
Shanxi 7365.7
Jilin 7203.2
Heilongjiang 8288.0
Anhui 10052.9
Jiangxi 7589.2
Henan 19367.3
Hubei 12831.5
Hunan 12930.7
Chongqing 6528.7
Sichuan 14151.3
Guizhou 3893.5
Yunnan 6168.2
Tibet 441.4
Shaanxi 8186.7
Gansu 3382.4
Qinghai 1081.3
Ningxia 1334.6
Xinjiang 4273.6
InnerMongolia 9725.8
Guangxi 7700.4

试试这个：
data tmp2;
id=_n_;
input region $ gdp;
cards;
Beijing 11865.9
Tianjin 7500.8
Hebei 17026.6
Liaoning 15065.6
Shanghai 14900.9
Jiangsu 34061.2
Zhejiang 22832.4
Fujian 11949.5
Shandong 33805.3
Guangdong 39081.6
Hainan 1646.6
Shanxi 7365.7
Jilin 7203.2
Heilongjiang 8288.0
Anhui 10052.9
Jiangxi 7589.2
Henan 19367.3
Hubei 12831.5
Hunan 12930.7
Chongqing 6528.7
Sichuan 14151.3
Guizhou 3893.5
Yunnan 6168.2
Tibet 441.4
Shaanxi 8186.7
Gansu 3382.4
Qinghai 1081.3
Ningxia 1334.6
Xinjiang 4273.6
InnerMongolia 9725.8
Guangxi 7700.4
;
proc sort;by gdp;
run;
%macro ranselect(data,method=,samprate=,n=,total=,out=,strata=, var=,export=0);
options nodate nonumber ps=100 ls=90;
%if &data ne %then %do;

%if &strata ne %then %do;proc sort data=&data;by &strata;%end;

%if &n ne %then %do;

proc surveyselect data=&data seed=20060704 method=&method n=&n out=&out;%end;

%if &samprate ne %then %do;

proc surveyselect data=&data seed=20060704 method=&method samprate=&samprate out=&out;%end;

%if strata ne %then %do;strata &strata;%end;

%if &var ne %then %do;id &var;%end;run;%end;
%else %do;
data aaaa;do id=1 to &total;output;end;run;
%if &strata ne %then %do;proc sort data=aaaa;by &strata;%end;
%if &n ne %then %do;
proc surveyselect data=aaaa seed=20060704 method=&method n=&n out=&out;%end;
%if &samprate ne %then %do;
proc surveyselect data=aaaa seed=20060704 method=&method samprate=&samprate out=&out;%end;
%if strata ne %then %do;strata &strata;%end;
%if &var ne %then %do;id &var;%end;run;
%end;
%if &export=0 %then %do;proc print data=&out noobs;run;%end;
%else %do;
proc export data=&out outfile='d:\tmp\ranselect.xls' dbms=excel2000 replace;%end;run;
ods select none;
proc datasets;delete &out;run;quit;ods select all;
%mend ranselect;
%ranselect(tmp2,method=srs,samprate=,n=20,total=31,out=bbbb,strata=,var=id region gdp,export=1);

wcguo94 发表于 2011-7-7 10:45
现需要做一项调查，要求按照变量GDP进行PPS抽样，要求：按照GDP大小成比例抽取20个地区样本，请问采用SAS如何实现？谢谢
Region GDP
Beijing 11865.9
Tianjin 7500.8
Hebei 17026.6
Liaoning 15065.6
Shanghai 14900.9
Jiangsu 34061.2
Zhejiang 22832.4
Fujian 11949.5
Shandong 33805.3
Guangdong 39081.6
Hainan 1646.6
Shanxi 7365.7
Jilin 7203.2
Heilongjiang 8288.0
Anhui 10052.9
Jiangxi 7589.2
Henan 19367.3
Hubei 12831.5
Hunan 12930.7
Chongqing 6528.7
Sichuan 14151.3
Guizhou 3893.5
Yunnan 6168.2
Tibet 441.4
Shaanxi 8186.7
Gansu 3382.4
Qinghai 1081.3
Ningxia 1334.6
Xinjiang 4273.6
InnerMongolia 9725.8
Guangxi 7700.4

There is a requirements for PPS method. In your example, the maximum size is 9. It cannot do n=20.

Hope the following example illustrate the idea clearly.

data TravelExpense;
length Region $20 GDP 8;
input Region GDP;
cards;
Beijing 11865.9
Tianjin 7500.8
Hebei 17026.6
Liaoning 15065.6
Shanghai 14900.9
Jiangsu 34061.2
Zhejiang 22832.4
Fujian 11949.5
Shandong 33805.3
Guangdong 39081.6
Hainan 1646.6
Shanxi 7365.7
Jilin 7203.2
Heilongjiang 8288.0
Anhui 10052.9
Jiangxi 7589.2
Henan 19367.3
Hubei 12831.5
Hunan 12930.7
Chongqing 6528.7
Sichuan 14151.3
Guizhou 3893.5
Yunnan 6168.2
Tibet 441.4
Shaanxi 8186.7
Gansu 3382.4
Qinghai 1081.3
Ningxia 1334.6
Xinjiang 4273.6
InnerMongolia 9725.8
Guangxi 7700.4
;
proc sql;
create table TravelExpense2 as
select *, sum(GDP) as TotalSize, GDP/sum(GDP) as RelativeSize, int(sum(GDP)/GDP) as max
from TravelExpense;
quit;
proc print;run;
proc surveyselect data=TravelExpense2 method=pps n=9 seed=123
out=sample20;
size GDP;
run;
proc surveyselect data=TravelExpense2 method=pps n=10 seed=123
out=sample20;
size GDP;
run;

学习啊，，，，，

学习了，谢谢啦

上述pps抽样中，n即抽取的省份不能超过9.也就是说31个中按照gdp规模pps抽样，最多抽9个。

http://www.mofeel.net/1169-comp-soft-sys-sas/11903.aspx