Normal Distribution Calculator

About Normal Distribution Calculator

How the Normal Distribution Formula Works

The normal (Gaussian) distribution is defined by two parameters: the mean (μ) and the standard deviation (σ). Its probability density function (PDF) is:

f(x) = (1 / (σ√(2π))) × e^{-(x - μ)² / (2σ²)}

To find probabilities, you convert your value to a z-score and then use the cumulative distribution function (CDF). The z-score formula is:

z = (x - μ) / σ

The CDF Φ(z) gives the probability that a standard normal variable falls below z. There is no closed-form expression for Φ(z), so it is computed using numerical approximations. This calculator uses the Abramowitz and Stegun rational approximation, which is accurate to about 7 decimal places (A&S formula 7.1.26, Handbook of Mathematical Functions, 1964).

Worked example: Suppose exam scores follow a normal distribution with a mean of 72 and a standard deviation of 8. What is the probability a student scores below 85?

1. Calculate the z-score: z = (85 - 72) / 8 = 1.625
2. Look up Φ(1.625) = 0.9479
3. There is a 94.79% chance a randomly chosen student scores below 85

If you need to find the raw z-score directly, the z-score calculator provides more detail on that conversion, including reverse lookups from probability to z-score.

One-Tailed vs Two-Tailed Probabilities

When working with normal distributions, the type of probability you need depends on the question you are asking. The distinction between one-tailed and two-tailed probabilities matters most in hypothesis testing, but it comes up any time you need to interpret an area under the bell curve.

A one-tailed probability measures the area in a single tail of the distribution. In this calculator, the "less than" mode gives the left-tail probability P(X < x), and the "greater than" mode gives the right-tail probability P(X > x). Use a one-tailed calculation when you have a directional question: "What fraction of parts weigh more than 52 grams?" or "What percentage of students scored below 60?"

A two-tailed probability measures the combined area in both tails, outside a symmetric range around the mean. The "between" mode in this calculator computes P(a < X < b), which is the area between two values. In hypothesis testing, two-tailed tests are used when you want to detect a difference in either direction. For example, testing whether a new drug changes blood pressure (up or down) rather than specifically lowers it.

The choice between one-tailed and two-tailed tests has practical consequences. A one-tailed test at α = 0.05 uses a critical z-value of 1.645, while a two-tailed test at the same significance level uses z = 1.96 (splitting the 5% rejection region across both tails, so 2.5% in each). This means a two-tailed test requires stronger evidence to reject the null hypothesis. Most published research uses two-tailed tests by convention, since specifying a direction before collecting data can introduce bias if the choice is not well-justified.

If you are running a formal hypothesis test and need the exact p-value from your test statistic, the p-value calculator handles both one-tailed and two-tailed conversions directly.

The 68-95-99.7 Rule and Why It Matters

The empirical rule (sometimes called the three-sigma rule) is a quick shortcut for normal distributions:

Range	Percentage of Data	Tail Probability (outside range)
μ ± 1σ	68.27%	31.73%
μ ± 2σ	95.45%	4.55%
μ ± 3σ	99.73%	0.27%
μ ± 4σ	99.9937%	0.0063%
μ ± 5σ	99.99994%	0.00006%

In quality control, the "six sigma" standard corresponds to 3.4 defects per million opportunities. The concept of sigma levels is fundamental to manufacturing, where companies use normal distribution statistics to set tolerance limits. According to the American Society for Quality, six sigma processes have been widely adopted across industries since the 1980s, with Motorola and General Electric being early pioneers.

The normal distribution appears so often in practice because of the central limit theorem (CLT). The CLT states that when you average a large number of independent random variables, the result approaches a normal distribution regardless of the original distribution of those variables. This is why sample means tend to be normally distributed even when the raw data is not.

For quick reference, here are some commonly used z-scores and their corresponding probabilities:

Z-Score	P(Z < z)	P(Z > z)	P(-z < Z < z)
0.5	0.6915	0.3085	0.3829
1.0	0.8413	0.1587	0.6827
1.5	0.9332	0.0668	0.8664
1.645	0.9500	0.0500	0.9000
1.96	0.9750	0.0250	0.9500
2.0	0.9772	0.0228	0.9545
2.326	0.9900	0.0100	0.9800
2.576	0.9950	0.0050	0.9900
3.0	0.9987	0.0013	0.9973

The z-scores of 1.645, 1.96, 2.326, and 2.576 are especially important in statistics. They correspond to the critical values for 90%, 95%, 98%, and 99% confidence levels respectively. If you are building confidence intervals from sample data, the sample size calculator uses these same critical values to determine how many observations you need for a given margin of error.

Common Applications of Normal Distribution Calculations

Normal distribution probabilities come up in a wide range of fields:

Education and testing: Standardised tests like the SAT, GRE, and IQ tests are designed so that scores follow a normal distribution. An IQ test, for instance, uses a mean of 100 and a standard deviation of 15. Scoring 130 puts you at the 97.7th percentile (z = 2.0). The standard deviation calculator can help you compute the spread of a dataset before using this tool.

Quality control: Manufacturing processes use normal distribution limits to define acceptable tolerances. If a factory produces bolts with a target diameter of 10mm and a standard deviation of 0.02mm, they might reject anything outside ±3σ (9.94mm to 10.06mm), accepting a 0.27% defect rate.

Finance: Short-term stock returns are often modelled as normally distributed for value-at-risk (VaR) calculations and option pricing (the Black-Scholes model assumes log-normal returns). A portfolio manager calculating the 5% VaR for a $1M portfolio with 2% daily standard deviation would look up z = -1.645 and find the worst expected daily loss is about $32,900.

Research and hypothesis testing: The normal distribution underpins most statistical tests. The widely used significance level of α = 0.05 corresponds to z = ±1.96 in a two-tailed test. If your test statistic exceeds 1.96, you reject the null hypothesis. For computing exact significance, the p-value calculator works directly from test statistics.

Medicine: Reference ranges for blood tests, growth charts, and drug response curves all rely on normal distribution assumptions. The World Health Organisation's growth standards for children are constructed using the LMS method, which transforms data to approximate normality.

Worked Example - Test Scores

A class of 200 students takes a final exam. The scores are normally distributed with a mean of 72 and a standard deviation of 10. Let's work through two practical questions using this distribution.

Question 1: What percentage of students scored above 85?

1. Calculate the z-score: z = (x - μ) / σ = (85 - 72) / 10 = 1.3
2. Look up Φ(1.3) in the CDF. From the table above (or using this calculator): P(Z < 1.3) = 0.9032
3. The right-tail probability is P(Z > 1.3) = 1 - 0.9032 = 0.0968
4. About 9.68% of students scored above 85. In a class of 200, that is roughly 19 students.

Question 2: What score marks the top 10% of the class?

This is the reverse problem. Instead of starting with an x-value and finding a probability, you start with a probability and find the x-value.

1. The top 10% means you need the value where P(X > x) = 0.10, or equivalently P(X < x) = 0.90
2. Find the z-score where Φ(z) = 0.90. From a standard normal table, z = 1.282
3. Convert back to the original scale: x = μ + zσ = 72 + (1.282 × 10) = 84.82
4. A student needs to score about 85 or higher to be in the top 10% of the class

Notice how both questions are related. The cutoff for the top 10% lands very close to 85, which lines up with the 9.68% figure from Question 1. These two approaches - forward (x to probability) and reverse (probability to x) - are the two fundamental operations you can perform with normal distributions. This calculator handles the forward direction. For the reverse lookup, the z-score calculator can convert a target percentile back to a raw score.

This type of analysis is common in education. College admissions offices, for example, use percentile cutoffs to set minimum test score requirements. The SAT has a mean of roughly 1060 and a standard deviation of about 210 (based on College Board data). A score of 1400 corresponds to a z-score of (1400 - 1060) / 210 = 1.62, placing a student around the 94.7th percentile.

When the Normal Distribution Does Not Apply

Not all data is normally distributed. Here are situations where you should use a different approach:

• Count data (e.g. number of website visitors per hour) - use a Poisson distribution instead
• Heavily skewed data (e.g. income, house prices) - consider a log-normal or gamma distribution
• Bounded data (e.g. percentages between 0% and 100%) - a beta distribution may be more appropriate
• Rare events (e.g. earthquake magnitude) - extreme value distributions handle tail behaviour better
• Small samples from unknown distributions - use the t-distribution instead of the normal, since it accounts for extra uncertainty

A quick way to check normality is to plot a histogram of your data and see if it looks roughly bell-shaped, or run a Shapiro-Wilk test. For small samples where the t-distribution is more appropriate, the confidence interval calculator automatically uses the correct distribution based on sample size.

You can also check for normality by calculating skewness and kurtosis. A normal distribution has a skewness of 0 (perfectly symmetric) and a kurtosis of 3 (sometimes reported as excess kurtosis of 0). Values far from these suggest your data may not follow a normal distribution closely.

Sources

• Abramowitz and Stegun - Handbook of Mathematical Functions, formula 7.1.26 (CDF approximation)
• NIST/SEMATECH e-Handbook of Statistical Methods - Normal Distribution
• American Society for Quality - Six Sigma overview and DPMO standards
• College Board - SAT Suite of Assessments Annual Report (score distributions)
• World Health Organisation - Child Growth Standards (LMS method)
• Investopedia - Normal Distribution in Finance and VaR