0
$\begingroup$

I have the folowing table:

enter image description here

according to this example, there are 40 observations distributed over 10 stores and 4 weeks of the month. Objective: to make a sample of 90%, 80%, 75% and 50% of the 40 observations so that: 1) as a result, all the stores were covered by the sample 2) as a result, all the weeks were covered by the sample. I'd call it a two-dimensional stratified sample, but I don't know what it's called.

$\endgroup$
0
$\begingroup$

Since you did not provide the data we will have to do this abstractly. We are talking about stratified sampling so you need to make sure that every cell gets the appropriate ratio of samples. Multiply each cell by 0.9 (for the first subsample) to get the number of samples that you need from each cell (plus minus rounding error). Then simply select 90 % of random samples from each cell. Together you will have roughly 90 % of observations from your original sample, for all shops and weeks.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.