Percentile: A value below which a given percentage of observations in a group of observations fall.
75th percentile (Q3): The value below which 75% of the data falls.
Quartiles: Dividing values that split the data into four equal parts (25th, 50th, 75th percentiles).
- Arrange data in ascending order
- Calculate the index: i = (p/100) × n, where p is the percentile and n is the number of values
- If i is an integer, percentile is the average of the i-th and (i+1)-th values
- If i is not an integer, round up to the next integer and take that value
Data is already arranged: 65, 70, 75, 80, 85, 90, 95, 100
Number of values (n) = 8
i = (75/100) × 8 = 0.75 × 8 = 6
Since i = 6 is an integer, take the average of the 6th and 7th values
6th value = 90
7th value = 95
75th percentile = (90 + 95) ÷ 2 = 185 ÷ 2 = 92.5
By definition, 75% of scores are below the 75th percentile
So 75% of scores are below 92.5
The 75th percentile is 92.5. This means 75% of the test scores are below 92.5, and 25% are above 92.5. The 75th percentile is also known as the third quartile (Q3).
• Index calculation: i = (p/100) × n
• Integer index: Average of i-th and (i+1)-th values
• Non-integer index: Round up and take that value
Z-score: The number of standard deviations a data value is from the mean.
Standardization: Converting raw scores to z-scores to compare values from different distributions.
Mean (μ) = 80
Standard deviation (σ) = 5
Raw score (x) = 90
z = (x - μ) ÷ σ
z = (90 - 80) ÷ 5
z = 10 ÷ 5 = 2
z = 2 means the score of 90 is 2 standard deviations above the mean
This is a relatively high score compared to the class average
|z| > 2 is typically considered unusual
A z-score of 2 indicates a score in the top 2.5% of a normal distribution
Score = Mean + (z × Standard deviation)
90 = 80 + (2 × 5) = 80 + 10 = 90 ✓
The z-score is 2. This means the student's score of 90 is 2 standard deviations above the mean of 80. The score is relatively high compared to the class, falling in the top portion of the distribution. A z-score of 2 indicates the score is significantly above average.
• Z-score formula: z = (x - μ) ÷ σ
• Positive z-score: Value above the mean
• Negative z-score: Value below the mean
Relative performance: How well someone performs compared to others in their group, accounting for the difficulty and variability of the assessment.
Standardization: Converting scores to z-scores allows comparison across different distributions.
Raw score = 85, Mean = 80, Standard deviation = 5
z_A = (85 - 80) ÷ 5 = 5 ÷ 5 = 1
Raw score = 92, Mean = 90, Standard deviation = 2
z_B = (92 - 90) ÷ 2 = 2 ÷ 2 = 1
Student A: z = 1
Student B: z = 1
Both students have the same z-score of 1
Despite different raw scores and different distributions, both students performed equally well relative to their respective classes
Both are 1 standard deviation above their class means
A z-score of 1 corresponds to approximately the 84th percentile
Both students performed better than about 84% of their respective classes
Both students performed equally well relative to their respective classes. Student A's score of 85 is 1 standard deviation above the class mean of 80, giving z = 1. Student B's score of 92 is 1 standard deviation above the class mean of 90, giving z = 1. Both students performed better than approximately 84% of their respective classes.
• Relative comparison: Z-scores allow comparison across different distributions
• Equal z-scores: Same relative performance
• Percentile conversion: Z-score of 1 ≈ 84th percentile
p% of data falls
Standard deviations from mean
Q2 = 50th %
Q3 = 75th %
Percentiles: Values that divide a rank-ordered set of data into 100 equal parts.
Z-scores: Standardized scores that indicate how many standard deviations a value is from the mean.
Standardization: Converting raw scores to z-scores to compare across different distributions.
- Percentile: Find value below which p% of data falls
- Z-score: Use formula (raw score - mean) ÷ standard deviation
- Interpretation: Positive values above mean, negative below mean
• Normal distribution: z = 1 ≈ 84th percentile, z = 2 ≈ 97.5th percentile
• Standardization: Transforms any normal distribution to standard normal (mean=0, std dev=1)
• Comparison: Higher z-score indicates better relative performance
Percentile rank: The percentage of values in a dataset that are less than a given value.
Rank-based percentile: Calculated using the formula: (rank - 0.5) ÷ n × 100%
Total number of values (n) = 20
Rank of the value = 4th position
Percentile rank = (rank - 0.5) ÷ n × 100%
Percentile rank = (4 - 0.5) ÷ 20 × 100%
Percentile rank = 3.5 ÷ 20 × 100%
Percentile rank = 0.175 × 100% = 17.5%
The 4th value has a percentile rank of 17.5%
This means 17.5% of the values in the dataset are less than this value
Alternatively, this value is at the 17.5th percentile
It's the value below which 17.5% of the data falls
If there are 20 values total and 3 values are below the 4th value, then 3/20 = 15%
The formula adds 0.5 to account for the position of the value itself
The 4th value has a percentile rank of 17.5%. This means that 17.5% of the values in the dataset are less than this value. The value is positioned relatively low in the distribution, with 82.5% of values above it.
• Percentile rank formula: (rank - 0.5) ÷ n × 100%
• Interpretation: Percentage of values below the given value
• Positioning: Lower ranks indicate lower relative positions
Population parameters: Mean and standard deviation for the entire population.
Standardized score: A z-score that allows comparison to population norms.
Population mean (μ) = 70 inches
Population standard deviation (σ) = 3 inches
Individual height (x) = 76 inches
z = (x - μ) ÷ σ
z = (76 - 70) ÷ 3
z = 6 ÷ 3 = 2
A z-score of 2 means the man's height is 2 standard deviations above the population mean
This is significantly taller than average
For a normal distribution, z = 2 corresponds to approximately the 97.5th percentile
This means he is taller than about 97.5% of adult males in the population
This is a very tall man relative to the population
Only about 2.5% of men in the population are taller than him
The man's z-score is 2, meaning he is 2 standard deviations (6 inches) above the population mean height. This corresponds to approximately the 97.5th percentile, indicating he is taller than about 97.5% of adult males in the population. He is significantly taller than average.
• Z-score formula: z = (x - μ) ÷ σ
• Normal distribution: z = 2 ≈ 97.5th percentile
• Population comparison: Z-scores indicate relative standing
Percentile: A measure indicating the value below which a given percentage of observations in a group of observations fall.
Z-score: A standardized score that indicates how many standard deviations an element is from the mean.
Standardization: The process of converting raw scores to z-scores to enable comparison across different distributions.
Normal distribution: A bell-shaped distribution where data is symmetrically distributed around the mean.
- Percentile calculation: Arrange data in order, find position using index formula
- Z-score calculation: Apply the standardization formula
- Interpretation: Convert to percentiles or standard benchmarks
• Standardization: Converts any normal distribution to standard normal (μ=0, σ=1)
• Percentile conversion: z = 1 ≈ 84th percentile, z = 2 ≈ 97.5th percentile
• Relative comparison: Higher z-score indicates better relative performance
• Outlier detection: |z| > 2 or |z| > 3 indicates unusual values
Percentiles: Rank-based, show relative position in distribution
Z-scores: Standardized, show distance from mean in standard deviations
Standardization: Enables comparison across different distributions
Normal distribution: Specific percentile-z-score relationships
Analysis: Percentiles and z-scores provide different but complementary ways to understand relative position in a distribution.
- Percentiles: Show what percentage of data falls below a value
- Z-scores: Show how many standard deviations a value is from the mean
- Standardization: Allows comparison across different scales
- Normal distribution: Z-scores have predictable percentile relationships
- Applications: Performance comparison, outlier detection, standardization