Box plot: A graphical display that summarizes the five-number summary (minimum, Q1, median, Q3, maximum) and shows the distribution of data.
- Calculate five-number summary for each dataset
- Create box plots for visual comparison
- Compare centers (medians)
- Compare spreads (IQR and range)
- Compare shapes (symmetry, skewness)
- Identify outliers in each distribution
| Class | Min | Q1 | Med | Q3 | Max |
|---|---|---|---|---|---|
| Class A | 65 | 72.5 | 82.5 | 92.5 | 100 |
| Class B | 60 | 67.5 | 77.5 | 87.5 | 95 |
Ordered: 65, 70, 75, 80, 85, 90, 95, 100
Min = 65, Max = 100
Median = (80 + 85) ÷ 2 = 82.5
Q1 = (70 + 75) ÷ 2 = 72.5
Q3 = (90 + 95) ÷ 2 = 92.5
Ordered: 60, 65, 70, 75, 80, 85, 90, 95
Min = 60, Max = 95
Median = (75 + 80) ÷ 2 = 77.5
Q1 = (65 + 70) ÷ 2 = 67.5
Q3 = (85 + 90) ÷ 2 = 87.5
Class A median: 82.5
Class B median: 77.5
Class A has higher median (better performance)
Class A IQR: 92.5 - 72.5 = 20
Class B IQR: 87.5 - 67.5 = 20
Both classes have same spread in middle 50%
Class A performed better on average (median = 82.5 vs 77.5)
• Center comparison: Compare medians for center differences
• Spread comparison: Compare IQRs for variability differences
• Visual analysis: Box plots allow quick comparison of distributions
Histogram: A graphical display that shows the distribution of quantitative data using bars to represent frequency within specified intervals.
| Height Range | Team A Frequency | Team B Frequency |
|---|---|---|
| 65-67 | 0 | 1 |
| 68-70 | 2 | 2 |
| 71-73 | 2 | 1 |
| 74-76 | 2 | 1 |
| 77-79 | 1 | 1 |
| 80-82 | 1 | 1 |
Range: 80 - 68 = 12 inches
Mean: (68+70+72+74+76+78+80) ÷ 7 = 74 inches
IQR: Q3 - Q1 = 78 - 70 = 8 inches
Range: 81 - 65 = 16 inches
Mean: (65+68+70+72+75+78+81) ÷ 7 ≈ 72.7 inches
IQR: Q3 - Q1 = 78 - 68 = 10 inches
Range comparison: Team B (16) > Team A (12)
IQR comparison: Team B (10) > Team A (8)
Both measures indicate Team B has more variability
Team B's range extends further in both directions
Histogram shows Team B covers wider range of heights
Team B has more variability (Range = 16 vs 12, IQR = 10 vs 8)
• Spread measures: Compare range, IQR, and standard deviation
• Visual confirmation: Histograms show distribution shape and spread
• Multiple measures: Use several measures to confirm findings
Dot plot: A simple graphical display that shows the distribution of quantitative data using dots plotted above a number line.
Pattern: Peaks at 5 hours, decreases on both sides
Shape: Approximately symmetric, unimodal
Center: Around 5 hours
Pattern: More evenly distributed across range
Shape: Appears more uniform, possibly slightly right-skewed
Center: Around 4.5 hours
Class X: More concentrated around center (bell-shaped)
Class Y: More spread out, wider range
Class X has clearer mode at 5 hours
Class X: More consistent study habits
Class Y: More varied study habits
Class Y extends to both ends of the range
Class X is more symmetric with clear center, Class Y is more spread out
• Shape identification: Look for symmetry, skewness, and modality
• Visual patterns: Dot plots reveal clustering and gaps
• Comparative analysis: Compare peaks, spread, and symmetry
Distribution: The pattern of values that a variable takes and how often it takes those values.
Center: A typical value that represents the middle of a distribution.
Spread: A measure of how spread out the values in a distribution are.
Shape: The overall pattern of a distribution, including symmetry, skewness, and modality.
Comparative analysis: The process of examining two or more distributions to identify similarities and differences.
- Visual inspection: Examine data displays for immediate differences
- Center comparison: Compare means and medians
- Spread comparison: Compare ranges, IQRs, and standard deviations
- Shape comparison: Identify symmetry, skewness, and modality
- Outlier detection: Note unusual values in each distribution
- Conclusion drawing: Summarize key differences and similarities
Comprehensive comparison: A thorough analysis examining center, spread, shape, and outliers of multiple distributions.
| City | Mean | Median | Range | IQR | Std Dev |
|---|---|---|---|---|---|
| City A | 5.25 | 5.25 | 5.5 | 3.0 | 1.87 |
| City B | 4.33 | 4.25 | 8.0 | 3.0 | 2.36 |
City A mean: 5.25, median: 5.25
City B mean: 4.33, median: 4.25
City A has higher center values
City A range: 5.5, IQR: 3.0, std dev: 1.87
City B range: 8.0, IQR: 3.0, std dev: 2.36
City B has greater overall spread but same IQR
City A: No outliers (all values within fences)
City B: 9.0 may be an outlier (check with 1.5×IQR rule)
Lower fence: 2.75 - 1.5(3.0) = -1.75
Upper fence: 5.75 + 1.5(3.0) = 10.25
9.0 is within fences, so no outliers
City A: More symmetric, evenly spaced data
City B: Possibly right-skewed due to higher maximum value
City A has higher rainfall with less variability; City B has more spread
• Comprehensive analysis: Examine all aspects of distribution
• Multiple measures: Use several statistics to validate findings
• Visual confirmation: Box plots support numerical analysis
Real-world application: Using statistical methods to solve practical problems and make informed decisions.
| Metric | Method 1 | Method 2 |
|---|---|---|
| Mean | 85.0 | 80.0 |
| Median | 85.0 | 80.0 |
| Range | 20.0 | 40.0 |
| IQR | 12.0 | 25.0 |
| Std Dev | 6.71 | 12.91 |
Method 1 mean: 85.0, median: 85.0
Method 2 mean: 80.0, median: 80.0
Method 1 has higher center values
Method 1 range: 20.0, IQR: 12.0, std dev: 6.71
Method 2 range: 40.0, IQR: 25.0, std dev: 12.91
Method 1 shows more consistent performance
Method 1: More concentrated around higher values
Method 2: More spread out, includes both very low and very high scores
Method 1 appears more effective due to higher average scores and greater consistency
Method 2 has potential for high achievement but also risk of low performance
Method 1 is more effective (higher mean, lower variability)
• Effectiveness criteria: Higher center values indicate better performance
• Consistency measure: Lower spread indicates more reliable outcomes
• Practical interpretation: Consider both average performance and reliability
Distribution comparison: The process of analyzing two or more datasets to identify similarities and differences in their statistical properties.
Measures of center: Statistical measures (mean, median, mode) that represent the typical or central value of a dataset.
Measures of spread: Statistical measures (range, IQR, standard deviation) that indicate the variability or dispersion of data.
Distribution shape: The pattern of how data values are distributed, including symmetry, skewness, and modality.
Outliers: Data values that are significantly different from other values in the dataset.
- Visual inspection: Examine data displays (box plots, histograms, dot plots) for immediate differences
- Center comparison: Compare means and medians to identify differences in location
- Spread comparison: Compare ranges, IQRs, and standard deviations to assess variability
- Shape analysis: Examine symmetry, skewness, and modality of distributions
- Outlier identification: Note unusual values in each distribution
- Contextual interpretation: Consider the practical meaning of differences
- Conclusion drawing: Summarize key findings and implications
| Metric | Purpose | Resistance to Outliers | When to Use |
|---|---|---|---|
| Mean | Center of distribution | Low | Symmetric data without outliers |
| Median | Center of distribution | High | Skewed data or with outliers |
| Range | Total spread | Low | Quick estimate of spread |
| IQR | Middle 50% spread | High | Robust measure of spread |
| Standard Deviation | Average distance from mean | Low | Normal distributions, all data matters |
Mean comparison: Arithmetic average, sensitive to outliers, good for symmetric data.
Median comparison: Middle value, resistant to outliers, good for skewed data.
Mode comparison: Most frequent value, useful for categorical or discrete data.
Interpretation: Higher center indicates better performance or higher values on average.
Range: Difference between max and min, sensitive to outliers.
IQR: Difference between Q3 and Q1, resistant to outliers.
Standard deviation: Average distance from mean, uses all data points.
Interpretation: Lower spread indicates more consistency and reliability.
Symmetry: Both sides of distribution are mirror images.
Right skew: Tail extends to the right, mean > median.
Left skew: Tail extends to the left, mean < median.
Modality: Number of peaks (unimodal, bimodal, multimodal).
Mean sensitivity: Outliers significantly affect the mean.
Median resistance: Outliers have minimal effect on the median.
Spread measures: Range heavily affected, IQR more resistant.
Visual identification: Box plots and dot plots help spot outliers.
Questions & Answers
Question: When comparing distributions, how do I know which measure of center to use? Should I always compare both mean and median?
Answer: The choice depends on your data's characteristics:
Use the mean when:
- Your data is symmetric (bell-shaped)
- There are no significant outliers
- You want to account for every data point
- You need to perform further mathematical calculations
Use the median when:
- Your data is skewed (has a long tail)
- There are outliers that would distort the average
- You want a measure resistant to extreme values
- You want to represent a "typical" value
Best practice: Compare both when possible, especially when distributions might be skewed. The relationship between mean and median tells you about skewness:
- Mean ≈ Median: Symmetric distribution
- Mean > Median: Right-skewed distribution
- Mean < Median: Left-skewed distribution
Question: I'm confused about when to use range vs IQR to compare variability. Which one is better?
Answer: Both measure spread, but they serve different purposes:
Range:
- Calculated as: Maximum - Minimum
- Uses only two data points
- Highly sensitive to outliers
- Provides the total spread of all data
- Simple to calculate but limited in information
Interquartile Range (IQR):
- Calculated as: Q3 - Q1 (middle 50% of data)
- Uses the middle half of all data points
- Resistant to outliers
- Focuses on the core of the distribution
- More robust measure of spread
Recommendation: Use IQR when comparing distributions, especially if outliers are present. Use range for a quick, rough idea of spread. IQR is generally preferred in statistical analysis because it's more reliable.
Question: How can I tell if a distribution is skewed just by looking at a box plot?
Answer: Box plots provide excellent visual clues for skewness:
Right skew (positive skew):
- The right whisker is longer than the left whisker
- The box is shifted to the left (closer to minimum)
- The median line is closer to the left side of the box
- The distance from median to Q3 is greater than from Q1 to median
Left skew (negative skew):
- The left whisker is longer than the right whisker
- The box is shifted to the right (closer to maximum)
- The median line is closer to the right side of the box
- The distance from Q1 to median is greater than from median to Q3
Symmetric distribution:
- Whiskers are approximately equal in length
- Box is centered in the plot
- Median is approximately centered in the box
- Distances from median to Q1 and Q3 are similar
Example: If the median is much closer to Q1 than to Q3, the distribution is right-skewed because the "tail" extends to the higher values.