# Assignment 3

## Due: Wednesday, October 21, 1998

In the textbook:

#3.3 parts a, c, d, and f

#3.15 give the p-value on part b rather than doing a 0.025-level test

#3.16 parts a, c, d, e, and f

#4.12

#4.13 part a

Also:

1. Study the TREES data that you used in Assignment 2. Regress volume on height and obtain the residuals and fitted values.

a) Plot the residuals against height, diameter, and the fitted values. Obtain a normal probability plot. Discuss what each plot says about the fit of the model.

b) Transform volume using the square-root of y, ln(y), and 1/y. Plot each transformed variable against height. Choose the transformation that makes the plot look most linear. What problem with lack of fit of the original model is resolved using this transformation?

c) Use the transformation you selected in part b) and regress it on height. Use this regression analysis and appropriate residual plots to assess the fit of your new model.

2. Again using the TREES data, regress volume on diameter. Obtain a Bonferroni joint confidence interval for B0 and B1 with family confidence coefficient of 0.95. Can you interpret this joint confidence interval by saying that 2.5% of the time such intervals will not capture B0 and 2.5% of the time such intervals will not capture B1? Explain.