You sell a given product through 1000 shops, and the average revenue for that product is K2.4 per week per shop, with a standard deviation of K3.1. You want to analyze if the offer "Three for the price of two!", in short 3x2, would be effective on your revenue. To do that, you introduce the offer in 50 shops for 1 week. Describe which data you would collect and what you would do with them. Would you recommend any special care in selecting the shops?

Since we do not know if the population is normally distributed (but we can assume that sample it is), should I use a t-Test? Utilizing stratified sampling could be useful inorder to cover theentire population(since shop revenue could change a lot among different regions)?