Investigate whether a transformation of the data in this case study might be desirable. Use the Box–Cox transformation on yield for several values of λ. You will need to add a small constant to avoid...

Investigate whether a transformation of the data in this case study might be desirable. Use the Box–Cox transformation on yield for several values of λ. You will need to add a small constant to avoid problems with the zero yields. Run PROC GLM (or a similar program) for each transformed yield variable and plot the residual sums of squares against λ. Construct the confidence interval on λ. What transformation is suggested?