k - when to include intercept?

[thesloth]

I thought K was the number of independent variables in a model….but I’ve seen a few examples that counted the intercept term when determining K. Help??

 

[S2000magician]

k is the number of independent variables; k + 1 is the number of regression coefficients you estimate (k slopes and 1 intercept).

 

[Xander086]

K is always the the number of slope coefficients. Remember Df degree of freedom = n - (k+1)
I think you meant Df when you want to include intercept ( the + 1).

 

[thesloth]

I have seen problems that for a standalone ‘k’ value that must include the intercept (there are four independent variables and one intercept, but the problem is solved using a ‘k’ value of five.) Meaning MSS = RSS/k, for example. Yet another error in the text??
Speaking of, random segue: in solving an interest tree problem, the solution starts with the value 108+8 rather than 100(par) + 8( coupon). Any idea where the heck that extra ‘8’ might have come from??
I am spending WAY too much time trying to separate errata from brain freeze, super frustrating!!