Question 1
You are given regression output of a model where the yearly salaries (in $) of CFL defensive players are regressed on their number of tackles performed in their career.
Salary=57298+155.26 Tackles
n=167 R2=0.145
(a) What would be the predicted salary of a really bad player?
(b) What is/are the advantages of running this model in log-log form?
(c) Interpret the coefficient in front of the regressor. Would you expect it to be in this range (negative/positive and size)?
(d) Do you believe that the coefficient represents the unbiased effect of an increase in the average number of tackles on the salary of players?
(e) Do you believe this is the correct specification of the model? Which changes, if any, would you make?
Question 2
You are regressing the number of kilometers travelled by car per week, a linear and random variable, on years of education and distance between the address of the individual and an arbitrarily determined geographic point in the closest city conisdered "downtown" for a sample of Canadians.
Do you expect to experience any problems with this regression?
Which ones? Justify and describe which elements (sign and symptoms, consequences observed in the model) would confirm/infirm your expectations.