What is an Inverse Normal Distribution Calculator?
An inverse normal distribution calculator works backwards from a probability to a value. Give it an area p and a normal distribution with mean μ and standard deviation σ, and it finds the cutoff x with P(X < x) = p, P(X > x) = p, or the pair x₁, x₂ for a central or two-tail area. It uses x = μ + σ · Φ⁻¹(area), the same calculation as invNorm on a TI-84.
What the Inverse Normal Distribution Calculator does
Choose whether your probability is a left tail, a right tail, a central area or two tails. Then enter the probability and the distribution’s mean and SD. The calculator returns the cutoff value (or pair of cutoffs), the matching z-score, a shaded curve, the invNorm call that gives the same answer, and a table of central intervals for the same distribution.
It is also called an invNorm calculator, an inverse norm calculator, a NORMINV calculator or an inverse normal CDF calculator. All of these names describe the same job: the inverse of the normal cumulative distribution function, also known as the normal quantile function or percent point function.
Typical questions are “what score is the 90th percentile?”, “what weight marks the heaviest 5%?”, “which range holds the middle 95% of values?” and “where should specification limits go so only 1% of parts fall outside?”. To go the other way, from a value to a probability, use the Normal Distribution Calculator. To turn a single value into a z-score, use the Z-Score Calculator. For critical values in hypothesis tests, the Z Critical Value Calculator is more direct.
How to use the Inverse Normal Distribution Calculator
- Pick the Area type that matches the wording of your question: “below” or “percentile” is a left tail, “above” or “top” is a right tail, “middle” is central, “outer” or “beyond” is two tails.
- Enter the Probability (area), p as a decimal (0.9) or a percentage (90%).
- Enter the Mean, μ and Standard deviation, σ. Use μ = 0 and σ = 1 for the standard normal, so the answer is the z-score itself.
- Read the cutoff value and z-score, and check the shaded curve to confirm the area is on the side you intended. The invNorm equivalent in the results shows the left-tail area to type into a TI-84 or spreadsheet.
Common inverse normal values (standard normal)
This short inverse normal table lists the z-scores used most often. Multiply z by σ and add μ to get the cutoff for your own distribution, x = μ + zσ. For any other area, use the calculator above or read the z table in reverse.
| Left-tail area, Φ(z) | Right-tail area | Central area (±z) | z = Φ⁻¹(area) |
|---|---|---|---|
| 0.80 | 0.20 | 0.60 | 0.8416 |
| 0.90 | 0.10 | 0.80 | 1.2816 |
| 0.95 | 0.05 | 0.90 | 1.6449 |
| 0.975 | 0.025 | 0.95 | 1.9600 |
| 0.99 | 0.01 | 0.98 | 2.3263 |
| 0.995 | 0.005 | 0.99 | 2.5758 |
| 0.999 | 0.001 | 0.998 | 3.0902 |
Inverse normal distribution formula
| Symbol | Meaning |
|---|---|
| p | the probability (area) you want, 0 < p < 1 |
| Φ⁻¹ | inverse standard normal CDF (quantile function) |
| μ, σ | mean and standard deviation of the normal distribution |
| x, x₁, x₂ | the cutoff value or values |
Every case reduces to one left-tail lookup, because the normal curve is symmetric: a right tail of p has 1 − p to its left, and a central area of p leaves (1 − p)/2 in each tail.
Φ⁻¹ has no closed form. Software uses a fast numerical approximation that is accurate to many decimal places, so results match a z-table to its printed precision.
Inverse normal worked example
Worked example: a 90th-percentile score
Problem. Exam scores are normal with μ = 500 and σ = 100. What score marks the 90th percentile?
- 01The 90th percentile has a left-tail area of 0.9.
- 02z = Φ⁻¹(0.9) = 1.2816.
- 03x = 500 + 1.2816 × 100 = 628.1552.
Answer. A score of about 628.16 is the 90th percentile. It is also the cutoff for the top 10%, because a right tail of p = 0.1 gives the same value. The 10th percentile sits the same distance below the mean, at 371.8448.
Inverse normal examples for each area type
The worked example above is a left tail. These three cover the other options. Each one runs in the calculator with the inputs shown.
| Question | Area type and inputs | z | Answer |
|---|---|---|---|
| IQ scores: what score marks the top 5%? | Right tail, p = 0.05, μ = 100, σ = 15 | 1.6449 | x = 124.67 |
| Exam scores: which range holds the middle 95%? | Central area, p = 0.95, μ = 500, σ = 100 | ±1.9600 | 304.00 to 696.00 |
| Fill weights: where do limits go so only 1% falls outside? | Two tails, p = 0.01, μ = 250, σ = 4 | ±2.5758 | below 239.70 or above 260.30 |
invNorm, NORM.INV and other software equivalents
Most software asks for the left-tail area only, so convert first: use 1 − p for a right tail and (1 + p)/2 for the upper end of a central area. These calls all give the worked example’s 628.1552. Leave out μ and σ (or use NORM.S.INV) to get the z-score 1.2816 instead.
Newer TI-84 Plus CE operating systems add a Tail setting (LEFT, CENTER or RIGHT) to invNorm, which does the conversion for you. Older models accept only a left-tail area.
| Tool | Function | Example |
|---|---|---|
| TI-83 / TI-84 | invNorm(area, μ, σ) | invNorm(0.9, 500, 100) |
| Excel | NORM.INV(p, μ, σ), or the older NORMINV | =NORM.INV(0.9, 500, 100) |
| Google Sheets | NORMINV(p, μ, σ) | =NORMINV(0.9, 500, 100) |
| R | qnorm(p, μ, σ) | qnorm(0.9, 500, 100) |
| Python (SciPy) | norm.ppf(p, μ, σ) | norm.ppf(0.9, 500, 100) |
| Python (standard library) | NormalDist(μ, σ).inv_cdf(p) | NormalDist(500, 100).inv_cdf(0.9) |
Common use cases
- Finding percentile cutoffs for test scores, growth charts or reference ranges.
- Setting specification or control limits that a given share of output should meet.
- Choosing a stock or capacity level that covers demand with a target probability (a service level).
- Finding the range that contains the middle 90%, 95% or 99% of a normal population.
- Finding z critical values such as 1.645 and 1.96 for confidence intervals and z-tests.
Assumptions and limitations
- The result is only as good as the normal model. For skewed data, empirical percentiles from the data are more trustworthy than μ + zσ.
- Extreme probabilities (very close to 0 or 1) depend heavily on the tails of the distribution, where real data most often depart from normality.
- Central and two-tail options assume the area is split equally between the two tails. For unequal splits, run two left-tail lookups instead.
- μ and σ are treated as known. If they are sample estimates, the cutoff itself is uncertain.
Frequently asked questions
How do I find x given a probability in a normal distribution?
Does invNorm give a z-score?
What is the difference between invNorm and normalcdf?
How do I find the top 10% cutoff of a normal distribution?
What range contains the middle 95% of a normal distribution?
Why is the z-score negative for a small left-tail probability?
Is invT the same as invNorm?
Cite this page
Statculator. (2026, November 14). Inverse Normal Distribution Calculator. Retrieved November 14, 2026, from https://statculator.com/normal-distribution/inverse-normal-distribution-calculator/