# What is the name of this type of stratifed sampling?

I am not sure what this specific type is called so I will describe what I did in Excel. This was employee data:

1. Sample size is prescribed. Example: n = 100 (though the population N = 10000)
2. Sort entire population (N) by by two fields, for example Organization and Employee Type , and get percentages out of total by each of these stratum.
3. Apply these percentages to get a random sample that is representative of entire population. For example if 10% of the population comes from Organiziation/Employee type stratum "Operations/ Engineer" than you should have 10 "Operations/ Engineers" in your sample of 100.

By the way, I did ask a similar question in Super User but there I more specifically asked what package and function to use in R to do this.

• The most important question to answer is whether or not when you sampled from each stratum did you do it at random? Jul 6, 2012 at 22:27
• yes. I edited step 3 to reflect that. Jul 6, 2012 at 22:51

## 2 Answers

If you sampled at random from each strata this would be called stratified random sampling proportional to size. If you did not sample at random from each group it would just be stratified sampling proportional to size and the representativeness of the sample could be brought into question depending on how the subjects were selected in each stratum.

• Conjugate Prior and I were independently answering during the same time. Proportional allocation and sampling proportionalt to size are synomous terms that are used for this type of stratification. But the adjective random should only be included if the sampling within strata were random. Jul 6, 2012 at 22:34

Among stratified random sampling schemes this is the one that uses proportional allocation to strata.

Here is a good brief summary of the usual possibilities, including the selection of 100 as the sample size, with some R code at the end.