(30 pts)
A marketing researcher studied annual sales of a product
that had been introduced 10 years ago. The data are as
follows, where X is the year (coded) and Y is sales in
thousands of units:
X Y
0 98
1 135
2 162
3 178
4 221
5 232
6 283
7 300
8 374
9 395
a) Prepare a scatterplot of the data. Does a linear
relation appear appropriate here? Superimpose the
OLS line.
b) Use the Box-Cox procedure to investigate the effects
of transforming Y. Produce a plot of the scaled SSE as
a function of lambda.
What transformation would you use?
c) Using the transformed data, plot the data and the
estimated regression line. Does the regression line
appear to be a good fit to the transformed data?
d) Compare residual plots and normal QQ plots of the
residuals for both the raw and the transformed data.
Is there an obvious improvement?