I need to know how to construct an experiment for this scenario?

here is what I would do.

1. In order to determine cause and effect you will need some stores that accept the coupon and others that do not in the same demographic. If you select all stores in St Louis and Kansas City to be part of the coupon accepting group you will need to select other stores in similar cities to be part of the base group that do not accept the coupons. One way to do this would be to use existing data to find sales of tissue in various regions of the country that are similar to the sales in the two selected cities.

2. Once you have the base stores selected you collect data from the two promotions. The data will be sales figures.

3. With the data from the promotions you can compare same store sales of the same tissue from previous months and from previous years for the same months. You will do regression analysis to determine whether there is correlation of sales to the coupons using the same store data. Correlation is not causation however.

4. You will then do a t test between the coupon store sales and the base store sales for the same months to see if any correlation you found is statistically significant.

so happy that I found this topic already answered! Its like you read my thoughts!