Unit 2.3 – Statistics for Two Categorical Variables
Statistical Analysis of Categorical Variables:
Use two-way tables to find proportions, compare groups, and detect associations between categorical variables.
Use two-way tables to find proportions, compare groups, and detect associations between categorical variables.
🔑 Core Statistics with Two-Way Tables
- Joint Proportion: Fraction of total in a specific cell (joint probability)
- Marginal Distribution: Proportion for each category, ignoring the other variable
- Conditional Distribution: Proportion in one group, for a given value of the other variable
- Association: If the distribution of one variable differs by category of the other, there may be association
Key Proportion Formulas
\[
\text{Joint Proportion} = \frac{\text{Cell Count}}{\text{Grand Total}}
\]
\[
\text{Marginal Proportion} = \frac{\text{Row or Column Total}}{\text{Grand Total}}
\]
\[
\text{Conditional Proportion} = \frac{\text{Cell Count}}{\text{Row or Column Total}}
\]
📊 Example Table Analysis
| Gender | Loves Math | Does Not Love Math | Total |
|---|---|---|---|
| Female | 12 | 8 | 20 |
| Male | 9 | 11 | 20 |
| Total | 21 | 19 | 40 |
- Overall (Marginal) Proportion who love math: \( \frac{21}{40} = 0.525 \)
- Conditional Proportion (Females who love math): \( \frac{12}{20} = 0.60 \)
- Conditional Proportion (Males who love math): \( \frac{9}{20} = 0.45 \)
- Association? Yes, females more likely in this sample to love math.
🖼️ Visuals for Two Categorical Variables
- Segmented Bar Chart: Stack proportions for each group to show differences
- Clustered Bar Chart: Compare proportions side by side
💡 Tips for Interpreting and Reporting
- Be clear about denominator: joint = all, marginal = totals, conditional = by group
- When row/column percents differ, association likely!
- Report both actual counts and proportions/percents on FRQs
- If all conditional proportions are equal, variables are likely independent
- Visualize with segmented or clustered bar charts for clarity
❌ Common Mistakes
- Confusing joint, marginal, and conditional percentages
- Using wrong totals in denominators
- Not stating which variable is “conditioned on” in conditional percents
- Claiming causation — association does not always mean causation!
- Forgetting to compare conditional proportions to look for association
Summary:
Unit 2.3 covers how to use two-way tables and bar chart visuals to derive and compare marginal, joint, and conditional proportions—key for uncovering associations between categorical variables. Always use clearly labeled counts and properly chosen denominators!
Unit 2.3 covers how to use two-way tables and bar chart visuals to derive and compare marginal, joint, and conditional proportions—key for uncovering associations between categorical variables. Always use clearly labeled counts and properly chosen denominators!