A retailer records advertising spend x x\,x thousand pounds and weekly sales y y\,y thousand pounds for branches in two regions. The data form two distinct clusters. The fitted lines are y=2.4x+18y=2.4x+18y=2.4x+18 for region A and y=1.1x+42y=1.1x+42y=1.1x+42 for region B. In both regions the observed values of x x\,x lie between 5 and 20.
For each region, estimate weekly sales when advertising spend is £12 000.
Compare the effect of an additional £1000 of advertising spend in the two models.
Explain why fitting one regression line to all the branches could be misleading.
A branch in region A plans to spend £30 000. Explain why its predicted sales should be treated cautiously.