topic Re: Creating a synthetic dataset with a maximum weight from a population count in SAS Data Management

Creating a synthetic dataset with a maximum weight from a population count

Demographer — Tue, 22 Oct 2019 09:15:56 GMT

I want to create a synthetic data that is representative of a population by age, sex, region and education. The original table looks like this (see attached for the full file):

agegr	edu	SEX	region	Population
0	e1	0	AD_rural	2180000
0	e1	0	AD_urban	1084307
0	e1	0	AN_rural	9476
0	e1	0	AN_urban	5178
0	e1	0	AR_rural	58663
0	e1	0	AR_urban	13887
…	…	...	…	…
100	E6	1	WB_rural	23

In the synthetic dataset, the weight of each individual should not be higher than 10,000. This means that I should create 218 individuals having the first set of variable (agegr=0, edu=e1, sex=0, AD_rural), each of them having a weight of 10,000.

Thank you

Re: Creating a synthetic dataset with a maximum weight from a population count

PeterClemmensen — Tue, 22 Oct 2019 09:20:27 GMT

I don't see a weight variable?

Re: Creating a synthetic dataset with a maximum weight from a population count

Demographer — Tue, 22 Oct 2019 09:21:50 GMT

There is none. The weight is now 1 for all line, since it's a population count.

Re: Creating a synthetic dataset with a maximum weight from a population count

Kurt_Bremser — Tue, 22 Oct 2019 10:51:42 GMT

Does this do it?

data have;
input agegr edu $ SEX :$1. region :$15. Population;
datalines;
0 e1 0 AD_rural 2180000
0 e1 0 AD_urban 1084307
0 e1 0 AN_rural 9476
0 e1 0 AN_urban 5178
0 e1 0 AR_rural 58663
0 e1 0 AR_urban 13887
;

data want;
set have;
weight = min(10000,population);
gross_weight = weight;
do while (gross_weight < population);
  output;
  weight = min(10000,population - gross_weight);
  gross_weight = gross_weight + weight;
end;
output;
keep agegr edu SEX region weight;
run;