The Data

We will begin with an example of simple logistic regression with a sample of the Western Collaborative Group Study. This study began in 1960 with 3154 men ages 39-59, who were employed in one of 11 California based companies. They were followed until 1969 during this time, 257 of these men developed coronary heart disease (CHD). You can read this data in with the code below. Lab 4 Markdown

Reading in the Data

library(haven)
wcgs <- read_dta("wcgs2.dta")
wcgs <- wcgs[, -16]

The Variables

Name Description
id Subject identification number
age Age in years
height Height in inches
weight Weight in lbs.
sbp Systolic blood pressure in mm
dbp Diastolic blood pressure in mm Hg
chol Fasting serum cholesterol in mm
behpat Behavior
1 = A1
2 = A2
3 = B3
4 = B4
ncigs Cigarettes per day
dibpat Behavior
1 = type A
2 = type B
chd69 Coronary heart disease
1 = Yes
0 = no
typechd Type of CHD
1 = myocardial infarction or death
2 = silent myocardial infarction
3 = angina perctoris
time169 Time of CHD event or end of follow-up
arcus Arcus senilis
0 = absent
1 = present
bmi Body Mass Index

  1. Begin by exploring the data and relationships further. What kind od relationships do you notice with out outcome of Coronary Heart Disease (CHD)? Make sure to use appropriate tables, graphs or summaries.

  2. Create logistic regressions with the following variables:
    • age
    • bmi
    • sbp
    • dbp
    • behpat
    • ncigs
    • dibpat
    • arcus Create a table with these univariate regressions. Interpret them.
  3. Consider a multiple with all of the above variables.
    1. Comment on the changes you see in coefficients from the univariate to multiple regression model. b. Pick the 3 most significant coefficients and interpret them.

  4. What is the log odds of CHD for a 65 year old with bmi of 36, sbp of 136, dbp 78, behpat A2, ncigs=3, dibpat Type B and No Arcus.

  5. What is the probability of CHD for a 65 year old with bmi of 36, sbp of 136, dbp 78, behpat A2, ncigs=3, dibpat Type B and No Arcus.