WELCOME TO

REGRESSION ANALYSIS IN PUBLIC HEALTH RESEARCH

JUNE 21-July 2, 2003

Course Objectives:

1) To improve students' ability to conduct regression analyses to study the relationship of a response variable on multiple predictor variables.

 

2) To sharpen the students skills in choosing a regression model, and interpreting regression coefficients.

 

3) The students should feel "comfortable" applying the sharpened skills to continuous responses (linear regression), binary responses (logistic regression) and time-to-event responses (survival analysis).

 

Course Materials:

 

Materials will be distributed in class, pdf files of the course materials are provided below:

 

Topic 1:  Multiple Linear Regression

Supplemental Handouts:    Standard output from linear regression model fit    

                                        Solution to the "You Do" on page 28

                                        Some more work with linear splines

                                        Solution to "Some more work with linear splines"

                                        More topic 1 notes

                                        R-square example

                                        Stepwise-selection example

                                        Testing of multiple regression coefficients

                                        More complicated modeling example

                                        Solution to "More complicated modeling example"

                                        Sample Size and Power Considerations

                                        Data for sample size and power considerations

 

Data from Topic 1 notes:  hospital data

 

Do Files for Topic 1 notes:     topic1v7.do

                                              topic1v8.do

 

Topic 2:  Logistic Regression

Supplemental Handouts:    Lab exercise 4

                                        Lab exercise 4 solution

                                        Rick Thompson's crossval.ado file for performing cross-validation

                                       

Do Files for Topic 2 notes:   topic2v7.do

                                           topic2v8.do

 

Data from Topic 2 notes:  kyphosis data

 

Topic 3: Survival Analysis

Supplemental Handouts:    Lab exercise 5

                                        Lab exercise 5 solution

                                        Lab exercise 6

                                        Lab exercise 6 solution

                                        Example of Poisson Regression 

 

Do Files for Topic 3 notes:  topic3v7.do

                                          topic3v8.do

                                          Example of Poisson Regression

 

 

 

PROJECT:  The project for both the Regression Analyses class and the Regression Analyses Lab are posted below.  In addition, the data file (project.dta) and the do-file (project.do) used to create all the figures for the Regression Analyses project file are posted below.

 

Project for Regression Analyses Course

Solution for Project for Regression Analyses Course

 

The data set:  project.dta

The do-file: project.do

 

 

Nice Reference Books for Regression Analyses:

 

Regression Modeling Strategies: With Applications to Linear Models, Logistic Regression, and Survival Analysis, Harrell, FE Jr. (2001). Springer-Verlag.

 

Applied Linear Regression, Weisberg, S. (1985).  2nd Ed. Wiley & Sons

 

Applied Logistic Regression, Hosmer and Lemeshow, 2nd Ed.  Wiley

 

Survival Analysis, Klein and Moeschberger, Springer-Verlag 

 

Statistical Methods in Medical Research, Second Edition.  Armitage, P. and Berry, G.  1987.  Blackwell Scientific, Oxford.  Chapters 10, 12 and 14.

 

Biostatistics: A Methodology for the Health Sciences.  Fisher, L. D. and vanBelle, G.  1993.  Wiley, New York, Chapters 11, 13, and 16.

 

Applied Regression Analysis and other Multivariable Methods.  Kleinbaum, D. G. and Kupper, L. L.  1978.    Duxbury Press, Boston.  Chapters 10, 12, 15 and 16.