01-20-2018 03:36 AM
I have a dataset wherein I have 1 lakh records based on data I need 300 samples of each based on one variable for example :
I have dataset which has data for India,Singapore,Malaysia
So I need 100 for India 100 for Singapore and 100 for Malaysia in one dataset . I cannot use where condition because I have more than 300 conditions for which I need 100 data in one dataset.
Any suggestion would really appreciate.
01-20-2018 04:20 AM
It sounds like you want simple random sampling by strata. The following SESUG paper describes this well (example 2 on page 7):