Homework:
Homework 1 -- due on 7 November.
You can see here an excellent job of solving Homework 1. The answers are all correct. Two minor comments. The graph for the menarche data should plot ln[p/(1-p)] on the y-axis rather than p itself. If you plot ln[p/(1-p)] vs. age you see that the relationship is linear. The Hosmer-Lemeshow test is useful when you have a numerical explanatory variable (like age or length of exposure) and you want to test whether the relationship to that variable is really linear. So it is a useful test for the neurology experiment and for the menarche data. But it is not useful for the final example, where the explanatory variables have only two categories.
Homework 2 -- due on 28 November.
Solution for Homework 2, Question 3.
Homework 3 -- due on 12 December.
The data file for Homework 3 with solutions
Written solutions for Homework 3.
Homework 4 -- due on 2 January.
The nasal cancer data file for Homework 4
The cranial growth data file for Homework 4, in a format suitable for repeated measures ANOVA
The cranial growth data file for Homework 4, in a format suitable for mixed model analysis
Homework 5 -- due on 23 January.
The data file for Hodgkins lymphoma. This is a plain text file.
The data file for Hodgkins lymphoma. This is an EXCEL file.
The data file for breast feeding. This is a plain text file and includes both the data and a description of the variables.
The data file for breast feeding. This is an EXCEL file.
The data file for bladder cancer. This is an EXCEL file.