Population Variance (σ²): The average of squared deviations from the mean for an entire population.
Population Standard Deviation (σ): The square root of population variance.
Formulas: σ² = Σ(xi - μ)²/N, σ = √σ²
- Calculate the population mean (μ)
- Find the deviation of each value from the mean
- Square each deviation
- Find the average of squared deviations (variance)
- Take the square root of variance (standard deviation)
Dataset: 5, 7, 9, 11, 13, 15
Sum = 5 + 7 + 9 + 11 + 13 + 15 = 60
N = 6 (total number of values)
μ = 60 ÷ 6 = 10
Deviations: (5-10), (7-10), (9-10), (11-10), (13-10), (15-10)
Deviations: -5, -3, -1, 1, 3, 5
Squared deviations: (-5)², (-3)², (-1)², (1)², (3)², (5)²
Squared deviations: 25, 9, 1, 1, 9, 25
Sum of squared deviations = 25 + 9 + 1 + 1 + 9 + 25 = 70
σ² = 70 ÷ 6 = 11.67
σ = √11.67 = 3.42
σ² = 11.67
σ = 3.42
Population Variance (σ²): 11.67
Population Standard Deviation (σ): 3.42
Interpretation: On average, values deviate by about 3.42 units from the mean.
• Population variance: σ² = Σ(xi - μ)²/N
• Population std dev: σ = √σ²
• Division by N: Use total population size (not N-1)
Sample Variance (s²): The sum of squared deviations divided by (n-1) to provide an unbiased estimate of population variance.
Sample Standard Deviation (s): The square root of sample variance.
Formulas: s² = Σ(xi - x̄)²/(n-1), s = √s²
Dataset: 8, 12, 15, 18, 22
Sum = 8 + 12 + 15 + 18 + 22 = 75
n = 5 (sample size)
x̄ = 75 ÷ 5 = 15
Deviations: (8-15), (12-15), (15-15), (18-15), (22-15)
Deviations: -7, -3, 0, 3, 7
Squared deviations: (-7)², (-3)², (0)², (3)², (7)²
Squared deviations: 49, 9, 0, 9, 49
Sum of squared deviations = 49 + 9 + 0 + 9 + 49 = 116
n-1 = 5-1 = 4 (degrees of freedom)
s² = 116 ÷ 4 = 29
s = √29 = 5.39
If treated as population: σ² = 116 ÷ 5 = 23.2
Sample variance (29) > Population variance (23.2)
Sample standard deviation (5.39) > Population std dev (4.82)
s² = 29
s = 5.39
Sample Variance (s²): 29
Sample Standard Deviation (s): 5.39
Interpretation: Sample uses n-1 divisor for unbiased estimation.
• Sample variance: s² = Σ(xi - x̄)²/(n-1)
• Sample std dev: s = √s²
• Division by n-1: Provides unbiased estimate of population variance
Degrees of Freedom: The denominator in variance calculation (N vs n-1) affects the result.
Bessel's Correction: Using n-1 instead of n to correct bias in sample variance estimation.
- Calculate population variance and standard deviation
- Calculate sample variance and standard deviation
- Compare the results
- Explain the theoretical basis for the difference
Dataset: 10, 12, 14, 16, 18
Mean (μ) = (10+12+14+16+18) ÷ 5 = 70 ÷ 5 = 14
Squared deviations: (10-14)², (12-14)², (14-14)², (16-14)², (18-14)²
Squared deviations: 16, 4, 0, 4, 16
Sum = 40
σ² = 40 ÷ 5 = 8
σ = √8 = 2.83
Using same dataset as sample: 10, 12, 14, 16, 18
Sample mean (x̄) = 14 (same as population mean)
Sum of squared deviations = 40
n-1 = 5-1 = 4
s² = 40 ÷ 4 = 10
s = √10 = 3.16
Population: σ² = 8, σ = 2.83
Sample: s² = 10, s = 3.16
Differences: Sample variance (10) > Population variance (8)
Differences: Sample std dev (3.16) > Population std dev (2.83)
The sample variance is larger because we divide by (n-1) instead of n
This is Bessel's correction to make sample variance an unbiased estimator of population variance
Sample statistics tend to be larger to account for sampling uncertainty
σ² = 8, σ = 2.83
s² = 10, s = 3.16
Population: σ² = 8, σ = 2.83
Sample: s² = 10, s = 3.16
Explanation: Sample variance is larger due to division by (n-1) instead of n.
• Population formula: Divide by N
• Sample formula: Divide by (n-1)
• Bessel's correction: Ensures unbiased estimation
Population Variance (σ²): The average of squared deviations from the population mean for an entire population.
Sample Variance (s²): The sum of squared deviations from the sample mean divided by (n-1), providing an unbiased estimate of population variance.
Population Standard Deviation (σ): The square root of population variance, measured in the same units as the original data.
Sample Standard Deviation (s): The square root of sample variance, providing an unbiased estimate of population standard deviation.
Bessel's Correction: The adjustment of using (n-1) instead of n in sample variance calculation to eliminate bias.
- Population variance: σ² = Σ(xi - μ)²/N
- Sample variance: s² = Σ(xi - x̄)²/(n-1)
- Population std dev: σ = √σ²
- Sample std dev: s = √s²
• Population variance: σ² = Σ(xi - μ)²/N
• Sample variance: s² = Σ(xi - x̄)²/(n-1)
• Standard deviation: σ = √σ², s = √s²
• Degrees of freedom: Use n-1 for sample to eliminate bias
Sampling considerations: Understanding when to treat data as sample vs population.
Practical implications: How the choice affects calculated statistics.
Dataset: 25, 28, 30, 32, 35, 38
Sample mean (x̄) = (25+28+30+32+35+38) ÷ 6 = 188 ÷ 6 = 31.33
Squared deviations: (25-31.33)², (28-31.33)², (30-31.33)², (32-31.33)², (35-31.33)², (38-31.33)²
Squared deviations: 40.44, 11.11, 1.77, 0.44, 13.44, 44.44
Sum = 111.64
s² = 111.64 ÷ (6-1) = 111.64 ÷ 5 = 22.33
s = √22.33 = 4.73
Using same dataset as population: 25, 28, 30, 32, 35, 38
Population mean (μ) = 31.33 (same as sample mean)
Sum of squared deviations = 111.64
σ² = 111.64 ÷ 6 = 18.61
σ = √18.61 = 4.31
Sample: s² = 22.33, s = 4.73
Population: σ² = 18.61, σ = 4.31
Sample statistics are larger due to division by (n-1)
If trees are a sample from a larger forest, use sample statistics (s = 4.73)
If trees represent the entire forest, use population statistics (σ = 4.31)
Sample statistic accounts for sampling uncertainty
s² = 22.33, s = 4.73
σ² = 18.61, σ = 4.31
Sample Statistics: s² = 22.33, s = 4.73
Population Statistics: σ² = 18.61, σ = 4.31
Conclusion: Sample statistics are larger due to Bessel's correction.
• Context matters: Determine if data represents population or sample
• Sample correction: Use n-1 for unbiased estimation
• Practical application: Results depend on research design
Convergence property: As sample size increases, sample and population formulas converge.
Degrees of freedom impact: The effect of using (n-1) diminishes with larger samples.
Assume dataset with sum of squared deviations = 100
Population std dev: σ = √(100÷10) = √10 = 3.16
Sample std dev: s = √(100÷9) = √11.11 = 3.33
Difference: 3.33 - 3.16 = 0.17
Same sum of squared deviations = 100
Population std dev: σ = √(100÷50) = √2 = 1.41
Sample std dev: s = √(100÷49) = √2.04 = 1.43
Difference: 1.43 - 1.41 = 0.02
Same sum of squared deviations = 100
Population std dev: σ = √(100÷100) = √1 = 1.00
Sample std dev: s = √(100÷99) = √1.01 = 1.005
Difference: 1.005 - 1.00 = 0.005
n=10: Difference = 0.17
n=50: Difference = 0.02
n=100: Difference = 0.005
As n increases, the difference approaches zero
The ratio n/(n-1) approaches 1 as n increases
For large samples, s² ≈ σ² and s ≈ σ
For small samples, the difference is more pronounced
Diff = 0.17
Diff = 0.02
Diff = 0.005
n=10: Difference = 0.17
n=50: Difference = 0.02
n=100: Difference = 0.005
Conclusion: As sample size increases, sample and population statistics converge.
• Sample size effect: Difference diminishes with larger samples
• Convergence: s → σ as n → ∞
• Practical implication: For large samples, difference becomes negligible
Variance: A measure of how spread out values are from the mean. It's the average of squared deviations from the mean.
Standard Deviation: The square root of variance, expressed in the same units as the original data, providing an intuitive measure of spread.
Population Parameters: σ² and σ represent parameters calculated from entire populations.
Sample Statistics: s² and s represent statistics calculated from samples to estimate population parameters.
Degrees of Freedom: The number of independent pieces of information used in calculation, typically n-1 for sample variance.
- Identify data type: Determine if population or sample
- Calculate mean: Population mean (μ) or sample mean (x̄)
- Find deviations: Subtract mean from each data point
- Square deviations: Eliminate negative values
- Sum squared deviations: Add all squared deviations
- Divide appropriately: By N for population, by (n-1) for sample
- Calculate standard deviation: Take square root of variance
• Population variance: σ² = Σ(xi - μ)²/N
• Sample variance: s² = Σ(xi - x̄)²/(n-1)
• Standard deviation: σ = √σ², s = √s²
• Degrees of freedom: Use n-1 for samples to eliminate bias
• Units: Standard deviation in same units as original data
Variance: A measure of statistical dispersion that quantifies how far each value in a dataset is from the mean. It's calculated as the average of the squared differences from the mean. Variance provides insight into the spread of data points around the central value.
Standard Deviation: The square root of variance, representing the average distance of data points from the mean in the original units of measurement. It's the most commonly used measure of spread because it's more interpretable than variance.
Population Parameters: Statistical measures calculated from an entire population (σ² and σ), representing the true values for the entire group being studied.
Sample Statistics: Statistical measures calculated from a subset of the population (s² and s), used to estimate population parameters with Bessel's correction applied.
Bessel's Correction: The adjustment of using (n-1) instead of n in sample variance calculation to eliminate bias and provide an unbiased estimate of population variance.
- Data identification: Determine if you have population or sample data
- Mean calculation: Calculate population mean (μ) or sample mean (x̄)
- Deviation calculation: Find difference between each value and the mean
- Squaring deviations: Square each deviation to eliminate negative values
- Summation: Add all squared deviations together
- Division: Divide by N (population) or (n-1) (sample)
- Standard deviation: Take square root of variance
- Interpretation: Analyze results in context
• Identify data type first: Determine if you have population or sample data before calculating
• Use correct formula: Population (÷N) vs Sample (÷(n-1))
• Understand Bessel's correction: Why we use n-1 for samples
• Interpret appropriately: Standard deviation is more intuitive than variance
• Consider sample size: Large samples make population/sample difference negligible
Questions & Answers
Question: Why do we divide by (n-1) instead of n when calculating sample variance? What is Bessel's correction?
Answer: Bessel's correction is a mathematical adjustment to make sample variance an unbiased estimator of population variance:
- Without correction: Sample variance would systematically underestimate population variance
- Reason: Sample mean is calculated from the same data, introducing dependency
- Correction: Dividing by (n-1) instead of n compensates for this bias
The sample mean is typically closer to the sample data points than the true population mean, which causes the squared deviations to be systematically smaller. Dividing by the smaller number (n-1) increases the variance to compensate for this underestimation.
This correction ensures that on average, our sample variance equals the population variance we're trying to estimate.
Question: Why do we square the deviations when calculating variance? Why not just use the absolute values?
Answer: Squaring deviations serves several important mathematical and statistical purposes:
- Mathematical convenience: Squared functions are differentiable and work well in calculus
- Emphasis on outliers: Squaring gives more weight to values far from the mean
- Positive values: Squaring ensures all deviations contribute positively to the measure
- Connection to normal distribution: Variance is fundamental to the normal distribution model
While mean absolute deviation (using absolute values) is a valid measure of spread, variance and standard deviation became standard because of their mathematical properties and their role in many statistical formulas and theorems.
The square root of variance (standard deviation) brings us back to original units while maintaining these beneficial properties.
Question: When is the difference between sample and population standard deviation significant? At what sample size does it become negligible?
Answer: The significance of the difference depends on the sample size:
- Small samples (n < 30): Difference is noticeable and important
- Moderate samples (30 ≤ n < 100): Difference is smaller but still present
- Large samples (n ≥ 100): Difference becomes negligible
The mathematical relationship is that the ratio n/(n-1) approaches 1 as n increases. For n=10, the ratio is 1.11; for n=100, it's 1.01; for n=1000, it's 1.001.
However, it's still important to use the correct formula regardless of sample size to maintain statistical rigor and consistency with standard practices.
The theoretical foundation (Bessel's correction) remains important even for large samples to ensure proper statistical inference.