Consider a true SLR data modelyi=0+1xi+i. If the error termsi are distributed i.i.dN(0,2), what can be said about the proposed transformed model yi=0+1/xi+i? To answer this, determine if the following SLR model assumptions hold after the transformation:

(a) Linearity of the relationship betweenXandY

(b) Pairwise independence ofi

(c) Homoscedasticity

(d) Normality ofi

(e) Independence ofifromxi

Are the answers to (a) through (e) the same for the transformed model 1/yi=0+1xi+i? If your answer is no, give a short explanation as to why.