11/9/10 Week 12
Association between Categorical Variables
What if study involved race and political party affiliation?
What if study involved analysis if relationship between smoking and lung cancer
Our class we will restrict our conversation to two categorical variables with only two levels for each
variable (ex. Gender, yes/no)
A university is concerned about financing it’s sports programs to include student input they are
interested to see if their exists a relationship between gender and attitude about equally supporting
male and female sports teams.
To answer: university takes a random sample from it’s student population and asks gender and yes
or no about supporting question.
1. What is gender? (Male/female) label this variable Gender
2. Believe in equal support? (Yes/no) label this variable equal support
Study is done where 877 students responded to study
From this we create a row by column rows “r” and “c” number of columns where r and c are simply
the number of levels for each respective variable.
We use as rows the levels for the explanatory categorical variable (ex. The variable that would
influence the other variable)
Gender No Yes Total
Female 72 466 538
Male 181 158 339
Total 253 624 877
Observed table (what we observed from sample data)
If there were no relation