I am looking for guidance on stratified sample sizes. I would like to analyse buying behaviour of businesses within a total population of 100,000.
There are 12 different types of businesses within the population. Each have different behaviours when it comes to what they buy, what volume they purchase, and how frequently.
I want to survey these businesses to analyse their different behaviour.
I know the total population, and the population of each business type.
I want to achieve high accuracy, but within a budget.
So my question is: what is the sample size?
Is it the statistically correct sample size for the total population, stratified by the business types?
Or is it the statistically correct sample size for each business type, added together?