Problem
An interesting way to extend the quadratic mean, random effects model for the Reisby data is to assume the following mean function:
μ(t) =
{β0 + β1t + β2t2 if endogenous
β′0 + β1t + β2t2 otherwise
That is, the two mean value curves are of the same shape. To put it still another way, moving one curve along the vertical axis gives the other curve.
• Write down the prediction equation for the last subject. Also, draw the prediction equations for the last ten subjects.