Comparaing populations with multiple variables

#1
My datasets represent the sales of a company with the phone calls they made to promote the products (some zones have sales but no phone calls). I also have the total population per zone which i Can add to my model (Ex: low sales because small population).
My objective is to determine:
  1. Does the phones calls increase the sales in general? North Vs South, In 2017, 2018 etc..
  2. Is there a significant difference between the south and the north zones?
The first table shows the geographic zones of the North of the country with the sales and the phone calls made in 2017 and 2018.

1551672830808.png

Idem second table but with the zones of the south.


1551672902379.png


Which(s) statistical analysis should I do to answer my questions?
 

obh

Active Member
#2
Hi Simon

I may think of multiple regression, you may think of Sales per population as DV.
But the number of observation seems to be too small to get significant results (before even checking if meeting the assumptions).

Is there a significant difference between the south and the north zones? yes because it isn't a sample ...You mean if the zone is good as a predictor for sales.

SalsPerPerson = 22.6464 - 15.2931 North - 1.2602 Calls - 8.4040 Year

It seems that phone calls were done when sales were low ...it doesn't say that phone calls are useless ...