Math
🆓 Free
Statistics
Standard Deviation Calculator
Paste any dataset and instantly get population SD (σ) or sample SD (s), variance, mean, median, range, coefficient of variation, standard error, and a complete step-by-step deviation table - plus a bell curve with 1σ/2σ/3σ bands and a built-in Z-score tool.
Enter Your Data
Z-Score Calculator
Calculate how many standard deviations a value is from the mean. First calculate SD in the Calculator tab, or enter values manually.
Z-Score Reference Table
| Z-Score | Percentile | Interpretation |
|---|
Standard Deviation Formulas
Population Standard Deviation (σ)
Use when you have data for the ENTIRE population.
σ = √(Σ(xᵢ - μ)² / N)
Where:
μ = population mean = Σx / N
N = total number of values
Σ(xᵢ - μ)² = sum of squared deviations
Where:
μ = population mean = Σx / N
N = total number of values
Σ(xᵢ - μ)² = sum of squared deviations
Sample Standard Deviation (s)
Use when your data is a SAMPLE from a larger population. Uses (n-1) - Bessel's correction.
s = √(Σ(xᵢ - x̄)² / (n-1))
Why n-1? Bessel's correction makes sample SD
an unbiased estimator of population SD.
With n-1, sample SD is slightly larger accounts
for the fact that a sample underestimates spread.
Why n-1? Bessel's correction makes sample SD
an unbiased estimator of population SD.
With n-1, sample SD is slightly larger accounts
for the fact that a sample underestimates spread.
Step-by-Step Process
1. Find mean: x̄ = Σx / n
2. Find deviations: (xᵢ - x̄) for each value
3. Square each: (xᵢ - x̄)²
4. Sum squares: Σ(xᵢ - x̄)²
5. Divide by n (pop) or n-1 (sample) Variance
6. Take square root Standard Deviation
2. Find deviations: (xᵢ - x̄) for each value
3. Square each: (xᵢ - x̄)²
4. Sum squares: Σ(xᵢ - x̄)²
5. Divide by n (pop) or n-1 (sample) Variance
6. Take square root Standard Deviation
Z-Score Formula
Z = (x - μ) / σ
Z tells you how many standard deviations
a value is above (+) or below (-) the mean.
|Z| < 1 within 1 SD (common)
|Z| 1-2 somewhat unusual
|Z| 2-3 unusual (only 4.5% of data)
|Z| > 3 very unusual (only 0.27% of data)
Z tells you how many standard deviations
a value is above (+) or below (-) the mean.
|Z| < 1 within 1 SD (common)
|Z| 1-2 somewhat unusual
|Z| 2-3 unusual (only 4.5% of data)
|Z| > 3 very unusual (only 0.27% of data)
The Empirical Rule (68-95-99.7)
In a normal distribution:
68.27% of data falls within μ ± 1σ
95.45% of data falls within μ ± 2σ
99.73% of data falls within μ ± 3σ
This is why anything beyond 3σ is called
an "outlier" - only 0.27% chance naturally!
68.27% of data falls within μ ± 1σ
95.45% of data falls within μ ± 2σ
99.73% of data falls within μ ± 3σ
This is why anything beyond 3σ is called
an "outlier" - only 0.27% chance naturally!
Coefficient of Variation (CV)
CV = (σ / |μ|) × 100%
CV measures relative variability.
Low CV data is tightly clustered
High CV data is widely spread
Useful for comparing variability across
datasets with different units or scales.
CV measures relative variability.
Low CV data is tightly clustered
High CV data is widely spread
Useful for comparing variability across
datasets with different units or scales.
Frequently Asked Questions
Standard deviation (SD) measures how spread out values in a dataset are from their mean. A low SD means values cluster tightly around the mean - data is consistent. A high SD means values are widely scattered - data is variable. It is the square root of variance, and it's the most widely used measure of spread in statistics, science, finance, and research.
Population SD (σ) divides by N and is used when your dataset contains every member of the group you're studying. Sample SD (s) divides by (n−1) - Bessel's correction - and is used when your data is a sample drawn from a larger population. The (n−1) prevents underestimating how spread out the full population actually is. For most real-world analysis - surveys, experiments, financial data - use sample SD. Use population SD only when you genuinely have data for every single member of the group.
Two reasons. First, deviations can be positive or negative - without squaring, they would cancel each other out and always sum to zero (since the mean is the balance point of the data). Squaring makes all values positive. Second, squaring gives more weight to values that are far from the mean, making SD more sensitive to outliers - which is usually what you want, since large deviations are disproportionately important in most real applications.
When you calculate SD from a sample, the sample mean is always closer to the sample data than the true population mean is. This causes the sum of squared deviations to be systematically smaller than it should be - leading to an SD that underestimates the true population spread. Dividing by (n−1) instead of n corrects for this bias. The result is an 'unbiased estimator' of population SD. The smaller your sample, the larger the correction - for n=2, it doubles the variance; for n=100, it barely changes it.
A Z-score measures how many standard deviations a value is from the mean: Z = (x − μ) / σ. Z = 0 means the value is exactly at the mean. Z = +1 means one SD above. Z = −2 means two SDs below. Z-scores are useful because they let you compare values across datasets with completely different units or scales. They also map directly to percentiles in a normal distribution - a Z-score of +1.65 corresponds to approximately the 95th percentile.
In a normal (bell curve) distribution: 68.27% of data falls within 1 SD of the mean (μ ± 1σ). 95.45% falls within 2 SD (μ ± 2σ). 99.73% falls within 3 SD (μ ± 3σ). Any value more than 3 SDs from the mean has only a 0.27% natural probability - this is why such values are called statistical outliers. Six Sigma quality control in manufacturing applies this logic at 6 SDs - targeting fewer than 3.4 defects per million opportunities.
CV = (SD / |mean|) × 100%. It expresses standard deviation as a percentage of the mean, allowing you to compare variability across datasets with different scales or units. A dataset with mean 1000 and SD 100 has CV = 10%. A dataset with mean 10 and SD 1 also has CV = 10% - same relative variability. CV is especially useful in finance (comparing stock volatility), biology (comparing measurement consistency), and quality control.
Standard deviation measures the spread of individual data points around the mean. Standard error (SE = SD / √n) measures how precisely your sample mean estimates the true population mean. As sample size increases, SE decreases - larger samples give more reliable mean estimates. SD stays roughly constant regardless of sample size (it describes the population spread). Use SD to describe variability in your data. Use SE when reporting how accurately your sample represents the broader population.
Standard Deviation Calculator - What It Is, How It Works, and When to Use It
Standard deviation is the most widely used measure of spread in statistics. It appears in school exams, scientific research, financial risk analysis, quality control, polling, and machine learning. Understanding what it means - not just how to calculate it - is what separates someone who can describe data from someone who can actually reason about it. This calculator does the full calculation with a step-by-step table so you can see exactly how each number is derived.
How to use this calculator: Type or paste your numbers separated by commas, spaces, or new lines. Select Population SD (σ) if you have data for every member of the group, or Sample SD (s) if your data is a subset. Click Calculate and get SD, variance, mean, median, range, CV, standard error, a deviation table, and a bell curve - all instantly.
Standard Deviation - The Core Idea
The mean tells you the centre of your data. Standard deviation tells you how far the data typically strays from that centre. Two datasets can have the exact same mean but completely different standard deviations - and that difference tells a completely different story.
- Dataset A: 49, 50, 51 Mean = 50, SD ≈ 0.82 (very tightly clustered)
- Dataset B: 10, 50, 90 Mean = 50, SD ≈ 32.7 (wildly spread out)
Same mean, completely different distributions. SD is what reveals which one you're dealing with.
Population SD vs Sample SD - Which One to Use
σ Population SD - Divide by N
- Use when you have data for the entire group (complete population)
- Examples: all test scores in a specific class, temperatures every day of one year, heights of all 30 students in a room
- Formula: σ = √(Σ(xᵢ − μ)² / N)
- Result is slightly smaller than sample SD
- No estimation needed - you have all the data
s Sample SD - Divide by (n−1)
- Use when your data is a sample drawn from a larger unknown population
- Examples: survey responses from 200 out of 10,000 customers, 50 product measurements from a production run, clinical trial participants
- Formula: s = √(Σ(xᵢ − x̄)² / (n−1))
- The (n−1) is Bessel's correction - prevents underestimating population spread
- This is the default for most real-world statistical analysis
The Step-by-Step Calculation - Exactly What the Table Shows
Every standard deviation calculation follows these six steps - the table in the results section shows you each one for your data:
- Find the mean (x̄): Add all values and divide by count. For [10, 20, 30]: mean = 60/3 = 20
- Find each deviation (xᵢ − x̄): Subtract the mean from each value. Results: −10, 0, +10
- Square each deviation (xᵢ − x̄)²: Makes all values positive and amplifies larger gaps. Results: 100, 0, 100
- Sum the squares Σ(xᵢ − x̄)²: Add all squared deviations. Total: 200
- Divide by N (population) or n−1 (sample): Get the variance. Population: 200/3 ≈ 66.67. Sample: 200/2 = 100
- Take the square root: Get standard deviation. Population SD ≈ 8.165. Sample SD = 10
What the Other Statistics Mean
The full summary this calculator provides goes beyond just SD:
- Variance: SD squared. Harder to interpret directly (it's in squared units) but mathematically useful in many formulas. SD = √Variance.
- Coefficient of Variation (CV): SD divided by mean, expressed as a percentage. Lets you compare variability across datasets with different scales. A stock with CV of 5% is far more stable than one with CV of 40%.
- Standard Error (SE): SE = SD / √n. Measures how precisely your sample mean estimates the true population mean. Smaller samples have larger standard error - more uncertainty about the true mean.
- Median: The middle value when sorted. When the median and mean are very different, your data is skewed - the normal distribution bands on the bell curve may not apply perfectly.
- Range: Max minus min. Quick but easily distorted by a single outlier - one extreme value blows up the range while SD remains informative.
The Empirical Rule - 68, 95, 99.7
For data that follows a normal distribution (the classic bell curve), the relationship between SD and the spread of data is precise and predictable:
- μ ± 1σ contains 68.27% of the data - roughly 2 out of 3 values
- μ ± 2σ contains 95.45% of the data - roughly 19 out of 20 values
- μ ± 3σ contains 99.73% of the data - nearly all values
This is why values beyond 3σ are called statistical outliers - there's only a 0.27% natural probability of landing there. In manufacturing, "Six Sigma" quality control targets defect rates below 6 standard deviations - roughly 3.4 defects per million.
Where Standard Deviation Is Used in the Real World
- Finance: Volatility of stocks and portfolios is measured as annualised SD of returns. Higher SD = higher risk. A fund with 20% annual SD is far more volatile than one with 5% SD.
- Education: Standardised test scores (SAT, IQ) are designed so that mean = 100 (or 500) and SD = 15 (or 100). Z-scores tell you exactly where a student ranks.
- Manufacturing / Quality Control: Six Sigma processes aim to keep defects within 6 standard deviations of the mean - 3.4 defects per million opportunities.
- Medicine: Clinical trial results are evaluated by how many standard deviations the treatment effect lies from zero. This is the basis of p-values and statistical significance.
- Weather forecasting: Forecast uncertainty is expressed as a range around the predicted value - often ± 1 SD of model ensemble spread.
- Machine learning: Feature normalisation (standardisation) converts values to Z-scores (mean=0, SD=1) so that features with different scales don't dominate model training.
How this calculator works, and where the numbers come from
The Standard Deviation applies the standard formula for this calculation to the values you enter and updates the result as you type. The calculation itself happens in your browser, and the page explains the method so you can check any result by hand.
Please note: Results are provided for general information and are calculated from the values you enter.