Relative frequency table: A table showing proportions or percentages instead of raw counts.
Joint relative frequency: The proportion of observations that fall into a specific combination of categories.
Marginal relative frequency: The proportion of observations for each category of one variable.
Conditional relative frequency: The proportion of observations in one category given another category.
- Start with the frequency table
- Divide each cell by the grand total
- Calculate row and column totals (marginal frequencies)
- Calculate conditional frequencies by dividing by relevant marginal totals
| Favorite Subject | Math | English | Total |
|---|---|---|---|
| Grade 9 | 30 | 20 | 50 |
| Grade 10 | 40 | 30 | 70 |
| Grade 11 | 25 | 35 | 60 |
| Total | 95 | 85 | 180 |
| Favorite Subject | Math | English | Total |
|---|---|---|---|
| Grade 9 | 0.167 | 0.111 | 0.278 |
| Grade 10 | 0.222 | 0.167 | 0.389 |
| Grade 11 | 0.139 | 0.194 | 0.333 |
| Total | 0.528 | 0.472 | 1.000 |
For Grade 9: P(Math|Grade 9) = 30/50 = 0.60, P(English|Grade 9) = 20/50 = 0.40
For Grade 10: P(Math|Grade 10) = 40/70 = 0.571, P(English|Grade 10) = 30/70 = 0.429
For Grade 11: P(Math|Grade 11) = 25/60 = 0.417, P(English|Grade 11) = 35/60 = 0.583
For Math: P(Grade 9|Math) = 30/95 = 0.316, P(Grade 10|Math) = 40/95 = 0.421, P(Grade 11|Math) = 25/95 = 0.263
For English: P(Grade 9|English) = 20/85 = 0.235, P(Grade 10|English) = 30/85 = 0.353, P(Grade 11|English) = 35/85 = 0.412
Joint frequencies show overall distribution: 16.7% are Grade 9 students who prefer Math
Marginal frequencies: 52.8% prefer Math, 47.2% prefer English
Conditional frequencies reveal patterns: Grade 11 students are more likely to prefer English (58.3%)
The relative frequency table shows that 52.8% of students prefer Math while 47.2% prefer English. However, conditional analysis reveals Grade-level differences: Grade 9 (60% Math), Grade 10 (57.1% Math), and Grade 11 (41.7% Math). This suggests an association between grade level and subject preference.
• Joint relative frequency: Cell value ÷ Grand total
• Conditional frequency: Cell value ÷ Relevant marginal total
• Sum property: Marginal frequencies sum to 1
Conditional probability: The probability of an event occurring given that another event has occurred.
Joint probability: The probability of two events occurring together.
Marginal probability: The probability of a single event occurring.
| Transportation | Car | Public Transit | Bicycle | Total |
|---|---|---|---|---|
| Young Adult (18-30) | 0.15 | 0.10 | 0.05 | 0.30 |
| Middle Age (31-50) | 0.20 | 0.15 | 0.05 | 0.40 |
| Senior (51+) | 0.25 | 0.05 | 0.00 | 0.30 |
| Total | 0.60 | 0.30 | 0.10 | 1.00 |
This is the conditional probability of preferring Car given the person is a Young Adult
Formula: P(Car | Young Adult) = P(Car and Young Adult) ÷ P(Young Adult)
P(Car | Young Adult) = 0.15 ÷ 0.30 = 0.50 or 50%
This is the conditional probability of being Young Adult given the person prefers Car
Formula: P(Young Adult | Car) = P(Car and Young Adult) ÷ P(Car)
P(Young Adult | Car) = 0.15 ÷ 0.60 = 0.25 or 25%
P(Car | Young Adult) = 0.50: Of all Young Adults, 50% prefer cars
P(Young Adult | Car) = 0.25: Of all car users, 25% are Young Adults
These probabilities answer different questions and have different values
Young Adults who prefer Car: 0.15 × 300 = 45 people
Total Young Adults: 0.30 × 300 = 90 people
45 ÷ 90 = 0.50 ✓
Young Adults are equally likely to prefer Car or Public Transit (50% vs 33.3%)
Seniors are most likely to prefer Car (83.3%)
Car is the most popular choice across all age groups
P(Car | Young Adult) = 0.50 means 50% of young adults prefer cars. P(Young Adult | Car) = 0.25 means 25% of car users are young adults. These probabilities address different questions: the first focuses on young adults' preferences, the second on the age composition of car users.
• Conditional probability: P(A|B) = P(A ∩ B) ÷ P(B)
• Order matters: P(A|B) ≠ P(B|A) unless P(A) = P(B)
• Interpretation: The condition determines the denominator
Independence: Two variables are independent if the occurrence of one doesn't affect the probability of the other.
Association: Two variables are associated if there's a relationship between their values.
| Transportation | Car | Public Transit | Bicycle | Total |
|---|---|---|---|---|
| Young Adult (18-30) | 0.15 | 0.10 | 0.05 | 0.30 |
| Middle Age (31-50) | 0.20 | 0.15 | 0.05 | 0.40 |
| Senior (51+) | 0.25 | 0.05 | 0.00 | 0.30 |
| Total | 0.60 | 0.30 | 0.10 | 1.00 |
Variables are independent if: P(A and B) = P(A) × P(B)
Or equivalently: P(A|B) = P(A)
P(Car) = 0.60
P(Young Adult) = 0.30
P(Public Transit) = 0.30
P(Senior) = 0.30
P(Car | Young Adult) = 0.15 ÷ 0.30 = 0.50
P(Car | Middle Age) = 0.20 ÷ 0.40 = 0.50
P(Car | Senior) = 0.25 ÷ 0.30 = 0.833
If independent: P(Car | Young Adult) = P(Car)
Actual: P(Car | Young Adult) = 0.50, P(Car) = 0.60
Since 0.50 ≠ 0.60, the variables are not independent
If independent: P(Car and Young Adult) = P(Car) × P(Young Adult)
Expected: 0.60 × 0.30 = 0.18
Actual: 0.15 ≠ 0.18
This confirms dependence between age and transportation preference
Age group and transportation preference are not independent. There is a significant association: seniors are more likely to prefer cars (83.3%) compared to young adults (50%) and middle-aged adults (50%). The conditional probabilities vary significantly across age groups, indicating that age influences transportation preference.
• Independence test: P(A|B) = P(A) for all combinations
• Joint probability test: P(A ∩ B) = P(A) × P(B)
• Association detection: Significant differences in conditional probabilities
P(A \cap B)
P(A) or P(B)
P(A|B) = P(A \cap B)/P(B)
Relative frequency: A proportion or percentage that represents how often an event occurs relative to the total number of observations.
Contingency table: A table that displays the frequency distribution of two categorical variables.
Conditional probability: The probability of an event given that another event has occurred.
Independence: Two events are independent if the occurrence of one does not affect the probability of the other.
- Convert to relative frequencies: Divide all cell values by grand total
- Calculate marginal frequencies: Row and column totals of relative frequencies
- Compute conditional frequencies: Divide cell values by relevant marginal totals
- Test for independence: Compare conditional to marginal probabilities
- Interpret associations: Analyze patterns in conditional frequencies
• Multiplication rule: P(A ∩ B) = P(A|B) × P(B)
• Sum of conditionals: Conditional probabilities in each row/column sum to 1
• Independence indicator: Large differences between conditional and marginal probabilities suggest association
Proportional transformation: Multiplying all frequencies by the same constant.
Non-proportional transformation: Adding a constant to all frequencies before conversion.
Original relative frequency: f/n
After doubling: (2f)/(2n) = f/n
Relative frequencies remain unchanged under proportional scaling
Original relative frequency: f/n
After adding 10: (f+10)/(n+k) where k is the total added to all cells
Relative frequencies change because the transformation is not proportional
Original table: A=10, B=20, Total=30 → Rel freq A=0.333, B=0.667
After doubling: A=20, B=40, Total=60 → Rel freq A=0.333, B=0.667 (unchanged)
After adding 5: A=15, B=25, Total=40 → Rel freq A=0.375, B=0.625 (changed)
Proportional transformations (multiplying by c) preserve relative frequencies
Non-proportional transformations change relative frequencies
This explains why sample size doesn't affect relative frequencies
But combining different samples changes the relative frequencies
When all frequencies are doubled, relative frequencies remain unchanged because the ratios stay the same. When a constant is added to each cell before conversion, relative frequencies change because the transformation is not proportional. This demonstrates that relative frequencies are preserved under proportional scaling but not under non-proportional transformations.
• Proportional scaling: Relative frequencies invariant under multiplication by constant
• Non-proportional scaling: Changes relative frequency distribution
• Ratio preservation: f/n = (cf)/(cn) for any constant c
Business analytics: Using statistical methods to analyze business data and inform decisions.
Customer segmentation: Dividing customers into groups based on shared characteristics.
| Satisfaction | New Customer | Loyal Customer | Total |
|---|---|---|---|
| Satisfied | 0.15 | 0.45 | 0.60 |
| Neutral | 0.10 | 0.15 | 0.25 |
| Dissatisfied | 0.10 | 0.05 | 0.15 |
| Total | 0.35 | 0.65 | 1.00 |
For New Customers: P(Satisfied) = 0.15/0.35 = 0.429, P(Neutral) = 0.10/0.35 = 0.286, P(Dissatisfied) = 0.10/0.35 = 0.286
For Loyal Customers: P(Satisfied) = 0.45/0.65 = 0.692, P(Neutral) = 0.15/0.65 = 0.231, P(Dissatisfied) = 0.05/0.65 = 0.077
For Satisfied: P(New) = 0.15/0.60 = 0.25, P(Loyal) = 0.45/0.60 = 0.75
For Neutral: P(New) = 0.10/0.25 = 0.40, P(Loyal) = 0.15/0.25 = 0.60
For Dissatisfied: P(New) = 0.10/0.15 = 0.667, P(Loyal) = 0.05/0.15 = 0.333
If independent: P(Satisfied | New) = P(Satisfied)
Actual: P(Satisfied | New) = 0.429, P(Satisfied) = 0.60
Since 0.429 ≠ 0.60, variables are dependent
Loyal customers are more satisfied (69.2%) than new customers (42.9%)
Dissatisfied customers are more likely to be new (66.7%) than loyal (33.3%)
There's a strong positive association between loyalty and satisfaction
There is a strong positive association between customer loyalty and satisfaction. Loyal customers are significantly more satisfied (69.2%) than new customers (42.9%). Conversely, dissatisfied customers are disproportionately new (66.7% of dissatisfied customers are new). This suggests that satisfaction drives loyalty, or that loyal customers develop higher satisfaction over time.
• Association detection: Compare conditional to marginal probabilities
• Business insight: Identify patterns for strategic decisions
• Dependency confirmation: Conditional probabilities differ from marginal
Relative frequency: The ratio of the number of times an event occurs to the total number of observations.
Joint relative frequency: The proportion of observations that belong to a specific combination of categories.
Marginal relative frequency: The proportion of observations that belong to a specific category of one variable, regardless of the other variable.
Conditional relative frequency: The proportion of observations in a specific category of one variable given that another variable has a specific value.
Independence: Two variables are independent if the conditional relative frequency equals the marginal relative frequency.
- Convert to relative frequencies: Divide each cell by the grand total
- Calculate marginal frequencies: Sum rows and columns
- Compute conditional frequencies: Divide by relevant marginal totals
- Test independence: Compare conditional to marginal frequencies
- Draw conclusions: Interpret the relationships found
• Joint = Marginal × Conditional: P(A ∩ B) = P(A) × P(B|A)
• Independence condition: P(A|B) = P(A) and P(B|A) = P(B)
• Sum property: Conditional probabilities in each row/column sum to 1
• Bayes theorem: P(A|B) = P(B|A) × P(A) ÷ P(B)
Joint frequencies: Proportion in specific cell
Marginal frequencies: Proportion for row/column
Conditional frequencies: Proportion within group
Independence: No association between variables
Association: Relationship between categorical variables
Analysis: Relative frequency tables convert raw counts to proportions, enabling comparison across different sample sizes.
- Joint frequencies: Show specific category combinations as proportions
- Marginal frequencies: Show individual variable distributions
- Conditional frequencies: Reveal relationships between variables
- Independence testing: Compare conditional to marginal probabilities
- Pattern identification: Look for meaningful associations