In the blood alcohol example using the least square regression line to make a prediction about blood alcohol content when drinking 20 beers 0 20 0 15 Blood alcohol content 0 10 0 05 0 0 0 1 2 3 4 5 Number of beers 6 7 7 8 would be an example of extended prediction O would help inform people about what blood alcohol content to expect on average O would help inform people about dangerous alcohol levels O would be improperly using the model for extrapolation Ta T 9