From time to time I will have customers ask me about the records that were selected when they used the Nth record sample option in the DeskTop Mailer and Business Edition software. So I will try to explain how this process works and how the software picks the records from the file.
Nth Record sampling is not random sampling. What it does is to take the sample size that you want and divide it into the total number of records in your file. This works very well until your sample size is greater than half of your file. For example: Say you have a file of 40 records and you want a sample of 4 records. The software would take 40 and divide it by 4 that would give you 10. So the software would select every tenth record from the file. If you wanted 5 records the software would take 40 divide it by 5 and then take every 8th record. The main question comes when you have the same file of 40 records but you want 30 in your sample. Because 30 is more than half of the records the software cannot take every X record and still get to 30. The software at this time will select the first 30 records leaving the last 10 records untouched. For samples that do not work out even the software will always round down. So for a file that has 50 records and you would like a sample of 23 that would divide down to 2.17 records so we would take every 2nd record until we get to the sample size of 23. For some application this is fine however for others this is not. A work around to this problem would be to take several smaller samples from the big file, each time exporting the selected records to a new file, and then deleting them from the original file. I hope this short explanation of Nth sampling is helpful to you.