Subject: Statistics I
Unit 4: Association Between Variables
Week 4 syllabus map
- L4.1: Association between two variables - Review of course
- L4.2: Association between two categorical variables - Introduction
- AQ4.2: Activity Question 2
- L4.3: Association between two categorical variables - Relative frequencies
- AQ4.3: Activity Question 3
- L4.4: Association between two numerical variables - Scatterplot
- AQ4.4: Activity Question 4
- L4.5: Association between two numerical variables - Describing association
- L4.6: Association between two numerical variables - Covariance
- AQ4.6: Activity Question 6
- L4.7: Association between two numerical variables - Correlation
- AQ4.7: Activity Question 7
- L4.8: Association between two numerical variables - Fitting a line
- AQ4.8: Activity Question 8
- L4.9: Association between categorical and numerical variables
How to use this week
- Separate the questions into three families:
- categorical with categorical
- numerical with numerical
- categorical with numerical
- For any association question, ask:
- what are the two variables?
- what type is each variable?
- what display is appropriate?
- are we describing direction, strength, or comparison across groups?
Week 4 exam traps
- Association does not automatically imply causation.
- Correlation measures linear association, not every possible relationship.
- Positive covariance and positive correlation point in the same direction, but their scales differ.
- A scatterplot may show no linear pattern even if a curved relationship exists.
- Relative frequencies are often more useful than raw counts when group totals differ.
Final revision checklist
- interpret association in words without overclaiming cause
- read two-way tables using counts and relative frequencies
- describe scatterplots by direction, form, and strength
- distinguish covariance from correlation
- explain what a fitted line is trying to do
- compare a numerical variable across categories sensibly