Multi stage sampling 

A method of sampling a population where you:

Divide into sub categories

One or more categories chosen

Take part samples of the categories chosen

(Also known as TWO stage cluster sampling)

To summarise think:

 Sampled samples 

i.e. sample categories of sample categories. 

To remember this think of the following:

A stage full of SAMPLES were given away at the beginning of the performance. You were given a sample of a sample.

Example 1

A multi-national bank has over 1000 branches spread over more than 20 countries and wants to conduct a survey of its customers to learn which services they like and dislike. How could they form a sample group comprising 1000 account holders using multi stage sampling?

Answer:

Stage 1 - Choose 5 countries the bank operates in using simple random sampling (or any other probability sampling).

Stage 2 – Choose 5 branches within each of the selected countries using systematic sampling method (or any other probability sampling).

Stage 3 – Choose 40 account holders to question from each branch using simple random or systematic sampling methods.

This will result in 40 (account holders) x 5 (Branches) x 5 (Countries) = 1000 customers