Do the previous problem, but use the football data on all NFL teams in the file P03_57.xlsx.
a. Rearrange the data so that there are six columns: Team, Year, Salary Last Year, Salary This Year, Wins Last Year, and Wins This Year. You don’t need rows for 2004 rows, because the data for 2003 isn’t available for Salary Last Year and Wins Last Year. Your ending data set should have 5*30 rows of data.
b. Run a multiple regression for Wins This Year versus the other variables (besides Team). Then run a forward stepwise regression with these same variables. Compare the two equations, and explain exactly what the coefficients of the equation from the forward method imply about wins.
c. The Year variable should be insignificant. Is it?
Why would it be contradictory for the “true” coefficient of Year to be anything other than zero?
d. Statistical inference from regression equations is all about inferring from the given data to a larger population. Does it make sense to talk about a larger population in this situation? If so, what is the larger population?