Normal Distribution Calculator

Background.

The Quanta normal distribution calculator answers the four questions that come up whenever you're working with a bell-shaped (Gaussian) distribution: what fraction of the population falls below a value, what fraction falls above it, what fraction falls between two values, and — in reverse — what value corresponds to a given probability. Enter a mean (μ) and a standard deviation (σ) to define your distribution N(μ, σ²), choose which of the four questions you're asking, and the calculator returns the answer along with the underlying z-score so you can see exactly how it was derived.

The normal distribution is the single most important probability distribution in applied statistics because so many real-world measurements — human heights, standardized test scores, manufacturing tolerances, measurement errors, and (by the Central Limit Theorem) the average of almost any large sample — cluster symmetrically around a central value in exactly this bell-shaped way. Its probability density function has no simple closed-form integral, so every practical calculation of 'what fraction of the distribution lies in this range' depends on a numerical approximation to the cumulative distribution function (CDF), traditionally looked up in a printed z-table and, in this calculator, computed directly using the Abramowitz-Stegun rational approximation — the same approximation used inside Excel, most calculators, and Quanta's own z-score calculator.

This calculator generalises that single-value z-score lookup into a full four-mode toolkot. The 'below' and 'above' modes answer the classic textbook question — what fraction of a normally distributed population scores below (or above) a given value — exactly what you'd look up in a z-table, but computed directly for any mean and standard deviation rather than requiring you to standardise by hand first. The 'between' mode answers a two-sided question that a single z-table lookup can't answer directly without an extra subtraction step: what fraction of the population falls in a specific range, such as the classic 'within one standard deviation' (68%), 'within two standard deviations' (95.45%), or any custom interval you specify. And the 'inverse' mode flips the whole problem around: instead of asking 'what's the probability for this value,' it asks 'what value corresponds to this probability' — the calculation behind every critical value you've ever seen quoted in a statistics class, from the z = 1.96 that anchors a 95% confidence interval to the cutoff score that defines the top 5% on a standardised test.

The worked default on this page uses the WAIS (Wechsler Adult Intelligence Scale) IQ scale, normed to a mean of 100 and a standard deviation of 15 — a scale chosen because its numbers are familiar and its results are easy to sanity-check by hand: an IQ of 130 is exactly two standard deviations above the mean, and the well-known fact that this sits at roughly the 97.7th percentile (in the top 2.3%) is a good way to confirm the calculator is doing exactly what you expect before you trust it on your own numbers. Change the mean, standard deviation, and mode to model anything else that plausibly follows a bell curve — birth weights, reaction times, measurement error in a lab instrument, or the sampling distribution of a sample mean under the Central Limit Theorem.

Every probability reported here uses the Abramowitz-Stegun §26.2.17 rational approximation to the CDF, accurate to better than 7.5×10⁻⁸ across the entire real line — effectively exact for any practical purpose — and the inverse mode uses the companion §26.2.23 approximation, accurate to within 4.5×10⁻⁴, the standard trade-off every statistics package makes because the inverse normal CDF has no closed form at all. If you only ever need to standardise a single value and read off its percentile, Quanta's dedicated z-score calculator is a faster, more focused tool for exactly that one job; this calculator is built for the fuller toolkit — arbitrary ranges, and the reverse direction — that a single z-score lookup can't provide.

What is normal distribution calculator?

The normal distribution (also called the Gaussian distribution, or informally the 'bell curve') is a continuous probability distribution defined by two parameters: a mean μ, which sets its centre, and a standard deviation σ, which sets its spread. Its probability density function is f(x) = (1 / (σ√(2π))) · exp(−(x−μ)²/(2σ²)), a symmetric bell shape that peaks at x = μ and falls off exponentially in both directions. The special case μ = 0, σ = 1 is called the standard normal distribution, and every other normal distribution can be converted to it by the standardisation z = (x − μ)/σ — the same transformation used throughout this calculator. Four quantities matter in practice, and this calculator computes all four. The cumulative distribution function Φ(z) = P(Z ≤ z) gives the probability of falling at or below a given standardised value; there is no elementary closed-form expression for this integral, so it must be approximated numerically — this calculator uses the Abramowitz-Stegun §26.2.17 rational approximation, accurate to 7.5×10⁻⁸. The complement, 1 − Φ(z), gives the probability of falling above a value. The difference Φ(z₂) − Φ(z₁) gives the probability of falling between two values. And the inverse, or percent-point function, Φ⁻¹(p), answers 'what z-value corresponds to cumulative probability p' — also with no closed form, approximated here via Abramowitz-Stegun §26.2.23, accurate to 4.5×10⁻⁴. Because so many natural and measured quantities are approximately normal — directly, or as a consequence of the Central Limit Theorem, which says the average of many independent random quantities tends toward a normal distribution regardless of the shape of the individual quantities — the normal distribution underlies confidence intervals, hypothesis tests, standardized test scoring, manufacturing quality control, and the great majority of introductory statistical inference.

How to use this calculator.

1. 01Enter the mean (μ) and standard deviation (σ) that define your distribution. Standard deviation must be strictly greater than zero.
2. 02Choose a mode: 'below' for P(X<x), 'above' for P(X>x), 'between' for P(a<X<b), or 'inverse' to find the x-value at a given probability.
3. 03In 'below'/'above' mode, enter the value x you're asking about.
4. 04In 'between' mode, enter a lower and upper bound (the lower bound must be strictly less than the upper).
5. 05In 'inverse' mode, enter a target probability strictly between 0 and 1 — for example 0.975 for the classic two-sided 95% confidence-interval cutoff.
6. 06Read the primary result, and check the z-value shown to see how far the relevant x-value sits from the mean in standard-deviation units.

The formula.

Every mode reduces to the same core computation: standardise, then apply the normal CDF (or its inverse). Standardisation converts an x-value on the distribution's own scale to a z-value on the standard normal scale: z = (x − μ) / σ. The standard normal PDF is φ(z) = (1/√(2π))·exp(−z²/2). The CDF Φ(z) = P(Z ≤ z) has no elementary closed form, so this calculator uses the Abramowitz & Stegun §26.2.17 rational approximation: for z ≥ 0, with k = 1/(1 + 0.2316419·z), Φ(z) ≈ 1 − φ(z)·(a₁k + a₂k² + a₃k³ + a₄k⁴ + a₅k⁵), using coefficients a₁ = 0.319381530, a₂ = −0.356563782, a₃ = 1.781477937, a₄ = −1.821255978, a₅ = 1.330274429; for z < 0, Φ(z) = 1 − Φ(−z) by symmetry. Maximum absolute error across the whole real line is 7.5×10⁻⁸ — indistinguishable from exact for any practical use. 'Below' mode returns Φ(z) directly. 'Above' mode returns 1 − Φ(z). 'Between' mode standardises both bounds and returns Φ(z_upper) − Φ(z_lower). 'Inverse' mode runs the process backwards: given a target probability p, it solves for the z that satisfies Φ(z) = p using the companion Abramowitz & Stegun §26.2.23 rational approximation for the percent-point function (also with no closed form), accurate to within 4.5×10⁻⁴, then converts back to the original scale via x = μ + σ·z. This pairing — a CDF approximation for the forward direction and a separate percent-point approximation for the inverse — is the standard approach used by essentially every statistics package, because Φ and Φ⁻¹ have no algebraic inverse relationship to exploit directly.

A worked example.

Frequently asked questions.

References& sources.

1. [1]Abramowitz, M., & Stegun, I. A. (Eds.). (1972). Handbook of Mathematical Functions (Dover, 9th printing). §26.2.17 — normal CDF rational approximation, |error| <= 7.5×10⁻⁸. §26.2.23 — inverse normal CDF (probit) rational approximation, |error| <= 4.5×10⁻⁴.
2. [2]NIST/SEMATECH e-Handbook of Statistical Methods, §1.3.6.6.1 — Normal Distribution. National Institute of Standards and Technology.
3. [3]NIST Digital Library of Mathematical Functions, Chapter 7 — Error Functions, Dawson's and Fresnel Integrals. Relates erf/erfc to the normal cumulative distribution function.
4. [4]Casella, G., & Berger, R. L. (2002). Statistical Inference (2nd ed., Duxbury). Chapter 3 covers the normal distribution family and its properties rigorously.
5. [5]OpenStax. Introductory Statistics 2e. Rice University, 2023. Chapter 6, 'The Normal Distribution' — worked examples and the empirical rule.
6. [6]WHO Multicentre Growth Reference Study Group. (2006). WHO Child Growth Standards. World Health Organization — a real-world application of normal-distribution z-scores at population scale.