Segmented Bar Graph: A bar graph where each bar is divided into segments representing subcategories, showing relative frequencies.
Relative Frequency: Proportion of each category within a group.
- Organize data in a two-way table
- Calculate relative frequencies for each group
- Create bars of equal height (100%)
- Divide each bar into segments based on relative frequencies
- Label segments and include a legend
freshmen: Bus = 30/50 = 0.60 (60%), Walk = 20/50 = 0.40 (40%)
sophomores: Bus = 40/50 = 0.80 (80%), Walk = 10/50 = 0.20 (20%)
juniors: Bus = 25/60 = 0.417 (41.7%), Walk = 35/60 = 0.583 (58.3%)
seniors: Bus = 20/40 = 0.50 (50%), Walk = 20/40 = 0.50 (50%)
Each bar will be 100% height with segments representing the relative frequencies.
Sophomores show highest bus usage (80%)
Juniors show highest walking rate (58.3%)
Seniors show equal distribution between bus and walking
Transportation preferences vary significantly by grade level.
Walking increases as students get older (from 20% to 58.3%).
The segmented bar graph reveals that transportation preferences differ significantly by grade level.
Sophomores rely most heavily on the bus (80%), while juniors prefer walking (58.3%).
• Relative Frequency: Each segment = (category count ÷ group total)
• Bar Height: All bars equal 100% for comparison
• Pattern Recognition: Compare segment sizes across bars
Association: A relationship between two variables where knowing the value of one variable provides information about the value of the other.
Independence: No relationship exists between variables.
Marketing: Desk = 30/50 = 0.60 (60%), Cafeteria = 20/50 = 0.40 (40%)
Sales: Desk = 40/70 = 0.57 (57%), Cafeteria = 30/70 = 0.43 (43%)
IT: Desk = 20/60 = 0.33 (33%), Cafeteria = 40/60 = 0.67 (67%)
Operations: Desk = 30/120 = 0.25 (25%), Cafeteria = 90/120 = 0.75 (75%)
Each bar represents a department with segments showing lunch preferences.
Marketing and Sales: Higher proportion eating at desk
IT and Operations: Higher proportion eating in cafeteria
Operations has the highest cafeteria usage (75%)
The conditional relative frequencies differ significantly across departments.
Knowing the department provides information about lunch location preference.
Therefore, there is an association between department and lunch location.
There is a clear association between department and lunch location preference.
Marketing and Sales employees prefer eating at their desks, while IT and Operations employees prefer the cafeteria.
Operations has the highest cafeteria usage (75%).
• Association Detection: Compare conditional relative frequencies
• Independence Test: Similar frequencies suggest independence
• Segmented Bar Graph: Equal height bars enable comparison
Multi-category Segmented Bar: Bars divided into multiple segments representing different categories.
Age Cohort: Group of people sharing similar age characteristics.
18-30: Smartphone = 80/150 = 0.533 (53.3%), Tablet = 40/150 = 0.267 (26.7%), Laptop = 30/150 = 0.200 (20.0%)
31-50: Smartphone = 60/150 = 0.400 (40.0%), Tablet = 50/150 = 0.333 (33.3%), Laptop = 40/150 = 0.267 (26.7%)
51+: Smartphone = 40/140 = 0.286 (28.6%), Tablet = 30/140 = 0.214 (21.4%), Laptop = 70/140 = 0.500 (50.0%)
Each bar represents an age group with three segments for device preferences.
18-30: Strong smartphone preference (53.3%)
31-50: Balanced preferences, slight smartphone preference (40%)
51+: Strong laptop preference (50%), lowest smartphone usage (28.6%)
Device preferences clearly vary by age group.
Younger people favor smartphones, older people favor laptops.
There is a strong association between age and device preference.
There is a strong association between age and device preference for online shopping.
Smartphone preference decreases with age (53.3% → 40% → 28.6%).
Laptop preference increases with age (20% → 26.7% → 50%).
• Multi-category Segments: Divide bars into multiple segments
• Pattern Recognition: Look for trends across categories
• Association Strength: Greater differences indicate stronger association
Segmented Bar Graph: A bar graph where each bar is divided into segments representing subcategories.
Relative Frequency: The proportion of each category within a specific group.
Association: A relationship between two categorical variables where knowing one variable provides information about the other.
Independence: No relationship exists between variables; conditional frequencies are similar across groups.
• Relative Frequency = Category Count ÷ Group Total
• Segment Size = Relative Frequency × 100% (for percentage bars)
• Association exists when conditional frequencies differ significantly
• Independence exists when conditional frequencies are similar
Income Level: Categorical variable with three levels (low, middle, high).
Living Arrangement: Categorical variable with three options (apartment, house, shared).
Low income: Apartment = 120/250 = 0.48 (48%), House = 80/250 = 0.32 (32%), Shared = 50/250 = 0.20 (20%)
Middle income: Apartment = 90/200 = 0.45 (45%), House = 100/200 = 0.50 (50%), Shared = 10/200 = 0.05 (5%)
High income: Apartment = 40/150 = 0.267 (26.7%), House = 110/150 = 0.733 (73.3%), Shared = 0/150 = 0.0 (0%)
Each bar represents an income level with segments for different housing types.
Low income: Highest shared housing usage (20%), balanced between apartment and house
Middle income: Nearly equal apartment and house, very little shared housing (5%)
High income: Overwhelming house preference (73.3%), no shared housing
There is a clear association between income and housing type.
As income increases, preference shifts from shared housing to houses.
High-income earners never choose shared housing.
There is a strong association between income level and housing choice.
Low-income earners prefer apartments (48%) and shared housing (20%).
High-income earners overwhelmingly prefer houses (73.3%) and never choose shared housing.
• Relative Frequency Calculation: Category count ÷ Group total
• Association Detection: Compare conditional frequencies across groups
• Pattern Recognition: Look for systematic changes in segment proportions
Clothing Style: Categorical variable with three options (casual, formal, athletic).
Gender: Categorical variable with two levels (male, female).
Male: Casual = 100/250 = 0.40 (40%), Formal = 80/250 = 0.32 (32%), Athletic = 70/250 = 0.28 (28%)
Female: Casual = 90/350 = 0.257 (25.7%), Formal = 110/350 = 0.314 (31.4%), Athletic = 150/350 = 0.429 (42.9%)
Each bar represents a gender with segments for different clothing styles.
Male: Highest casual preference (40%), followed by formal (32%) and athletic (28%)
Female: Highest athletic preference (42.9%), followed by formal (31.4%) and casual (25.7%)
Significant differences in style preferences between genders
There is a clear association between gender and clothing style preference.
Males prefer casual wear more than females (40% vs 25.7%)
Females prefer athletic wear more than males (42.9% vs 28%)
There is an association between gender and clothing style preference.
Males prefer casual wear significantly more than females (40% vs 25.7%).
Females prefer athletic wear significantly more than males (42.9% vs 28%).
Formal wear preferences are similar between genders.
• Relative Frequency Calculation: Each segment = Category count ÷ Group total
• Association Detection: Compare segment sizes across bars
• Pattern Analysis: Look for systematic differences in preferences
Complete Summary: Segmented Bar Graphs
Segmented Bar Graph: A bar graph where each bar is divided into segments representing subcategories.
Relative Frequency: Proportion of each subcategory within a group.
Conditional Relative Frequency: Frequency of subcategory given a specific group.
Equal Height Bars: All bars have same height (usually 100%) for comparison.
Comparison Tool: Enables comparison of subcategory proportions across different groups.
Association Detection: Different segment sizes across bars indicate association.
Visual Pattern Recognition: Easy identification of trends and differences.
Standardized Format: All bars equal 100% for fair comparison.
Step 1: Organize data in a two-way table format.
Step 2: Calculate relative frequencies for each group separately.
Step 3: Create bars of equal height (100%).
Step 4: Divide each bar into segments based on relative frequencies.
Step 5: Label segments and include a legend.
Association Present: Different segment proportions across bars.
Independence: Similar segment proportions across bars.Strength: Greater differences indicate stronger association.
Direction: Identify which groups prefer which categories.
Segmented Bar Graph: A graphical representation showing the distribution of a categorical variable within each category of another variable.
Relative Frequency: The proportion of observations in a particular category relative to the total number of observations in that group.
Conditional Relative Frequency: The relative frequency of a subcategory given a specific category of another variable.
Association: A relationship between two variables where the conditional relative frequencies differ across categories.
Independence: No relationship exists between variables; conditional relative frequencies are similar across categories.
- Data Organization: Arrange data in a two-way table.
- Relative Frequency Calculation: Calculate for each group separately.
- Graph Construction: Create segmented bars of equal height.
- Pattern Recognition: Compare segment sizes across bars.
- Association Assessment: Determine if variables are related.
• Relative Frequency = Category Count ÷ Group Total
• Segment Size = Relative Frequency × 100% (for percentage bars)
• Association exists when conditional frequencies differ significantly
• Independence exists when conditional frequencies are similar
Questions & Answers
Question: What's the difference between a regular bar graph and a segmented bar graph? When should I use each?
Answer: The key differences are:
Regular Bar Graph:
- Shows frequency or count for each category of ONE variable
- Each bar represents a single category
- Used for comparing magnitudes across categories
- Example: Number of students in each grade level
Segmented Bar Graph:
- Shows the distribution of ONE variable within categories of ANOTHER variable
- Each bar is divided into segments representing subcategories
- Used for comparing proportions across groups
- Example: Transportation method preferences by grade level
Use segmented bar graphs when you want to compare the composition or relative frequencies of subcategories across different groups.
Question: How do I know if there's an association between variables using a segmented bar graph? What exactly should I look for?
Answer: To detect association in segmented bar graphs:
Look For:
- Different segment proportions: If segments have different sizes across bars
- Systematic patterns: Consistent changes in segment sizes across groups
- Contrasting preferences: One group strongly favors one category while another favors a different category
Examples:
- Association: Young people prefer smartphones (large segment), old people prefer laptops (large segment)
- Independence: All groups have similar proportions of smartphone, tablet, and laptop users
If the conditional relative frequencies (segment sizes) differ significantly across the groups, there's an association. If they're similar, the variables appear independent.
Question: In Exercise 5, you said there's an association between gender and clothing style, but does that mean gender causes people to prefer certain styles?
Answer: This is a crucial distinction in statistics:
Association vs. Causation:
Association: Variables tend to occur together (correlation)
Causation: One variable directly causes changes in another
Confounding Variables:
Gender might correlate with other factors like social norms, lifestyle, occupation, or cultural expectations that influence clothing preferences. These could be the actual causes.
Establishing Causation:
- Requires controlled experiments
- Must eliminate confounding variables
- Needs to establish temporal order (cause before effect)
- Should show a dose-response relationship
We can say gender is associated with clothing style preferences, but we cannot conclude gender causes these preferences without a controlled experiment. The association might reflect cultural norms, social expectations, or other underlying factors.