Assignment 6

Due Monday, April 1.

Turn in problem # 1 (15 points).

Problem 1:

Return to the SMSA pollution data analyzed as an example in class. Replace the pollutant variables in the data frame by their logarithms. You may also transform others of the predictors if you feel that such transformations are called for. Then use variable selection techniques of your choice to identify a ``good'' model or models. Discuss these model(s). Your write up and discussion should include:

• Computation and interpretation of R-squared, and the Cp and PRESS statistics for the model(s) you discuss.
• Discussion/interpretation of coefficient estimates.
• Discussion of how closely competing models differ from one another.
• On the basis of your analysis, what can be said about the relationship between the pollutants and mortality? Can a causal link between pollution and mortality be established on the basis of this data? What can be said if anything?
• How does your analysis compare to the results obtained without transforming the pollutant variables?

Rawlings:

# 7.1, 7.2, 7.3, 7.4, 8.3, 8.5