InvNorm Calculator (Inverse Normal Distribution)
Enter a probability to get the z-score or x-value that cuts off that area under the normal curve. It works like invNorm(area, μ, σ) on a TI-84, and adds left, right, center and two-tail modes.
Enter the left cumulative probability P(X ≤ x).
Enter a probability to see results.
Enter a probability (area) and this invNorm calculator returns the z-score, or, with a mean and standard deviation, the x-value, that cuts off that area under the normal curve. It works like invNorm(area, μ, σ) on a TI-84: by default the area is to the left, and you can switch to right-tail, center or two-tail problems without converting by hand. For example, invNorm(0.95) = 1.64485. Each result comes with a shaded graph, the steps, and the same calculation in Excel, R and Python.
InvNorm in 30 seconds
What does invNorm do?
invNorm (inverse normal) takes an area under the normal curve, a probability, and returns the value that cuts off that area. By default the area is to the left, so invNorm(p) answers "which z-score has p of the distribution below it?" It is the reverse of normalCdf. Use it to find a z-score from a probability, a percentile cutoff or a critical value.
Does invNorm give a z-score or an x-value?
Both, depending on what you enter. invNorm(p) returns a z-score from the standard normal distribution (μ = 0, σ = 1). invNorm(p, μ, σ) returns an x-value on your data's scale, computed as x = μ + z·σ.
invNorm(0.95) and other values you'll use most
| Left-tail area p | invNorm(p) | Typical use |
|---|---|---|
| 0.90 | 1.28155 | 90th percentile |
| 0.95 | 1.64485 | one-tailed 5% cutoff |
| 0.975 | 1.95996 | two-tailed 95% critical value |
| 0.99 | 2.32635 | one-tailed 1% cutoff |
| 0.20 | −0.84162 | 20th percentile |
Select a value to load it into the calculator.
How to use this invNorm calculator
- Pick the area your problem describes: left tail, right tail, middle area or two tails.
- Choose the distribution: Standard Normal for a z-score, or Custom Normal to enter μ and σ and get an x-value.
- Enter the probability: 0.95, .95, 95 and 95% all work.
- Read the result: the z-score or x-value, the percentile, a shaded graph, the steps, and the matching TI-84, Casio, Excel, R and Python entry.
Left, right, center or two tails: which area to enter
On recent TI-84 Plus CE operating systems the full syntax is invNorm(area[, μ, σ, tail]), where tail is LEFT, CENTER or RIGHT. This calculator uses the same three options and adds a two-tails option:
- LEFT (default): the area is to the left of the cutoff, P(X ≤ x) = p. Use it for percentiles. Every older TI model and every spreadsheet function works this way.
- RIGHT: the area is to the right, P(X ≥ x) = p, for questions like "what value is exceeded by only 5%?" The calculator converts this to the left area 1 − p for you.
- CENTER (middle area): the area is the middle of the curve, symmetric around the mean. Enter 0.95 to get the bounds of a 95% interval.
- Two tails: the area is split equally between both ends. Enter a significance level α such as 0.05 to get both critical values.
Not sure which one you have? See left tail vs right tail probability.
Standard normal vs custom mean and standard deviation
Use Standard Normal when the question asks for a z-score or critical value. Use Custom Normal when it gives a mean and standard deviation (test scores, heights, measurements) and asks for a cutoff in those units. If you are given a variance instead, tick the variance option.
Reading your result: z, x and percentile
The z-score tells you how many standard deviations the cutoff sits from the mean. The x-value is that cutoff in your data's units. The percentile is the share of the distribution below the cutoff. A plain-language line under each result explains it in the form most homework answers expect, for example "95% of values fall below 1.64485."
No graphing calculator? This page works in any browser. You can also reverse-look-up a standard normal (z) table, though that is limited to 2 decimals.
InvNorm examples with solutions
Left tail: the 90th percentile
Test scores are normally distributed with mean 500 and standard deviation 100. What score marks the 90th percentile?
p = 0.90 (left tail). z = Φ−1(0.90) ≈ 1.28155, so x = 500 + 1.28155 × 100 ≈ 628.16. On a TI-84: invNorm(0.90, 500, 100).
Try this example in the calculator
Right tail: a top-5% cutoff
Bolt diameters are normally distributed with mean 10 mm and standard deviation 0.2 mm. What diameter is exceeded by only 5% of bolts?
Right tail p = 0.05, so the left area is 0.95. z ≈ 1.64485, so x = 10 + 1.64485 × 0.2 ≈ 10.32897 mm. On a TI-84: invNorm(0.95, 10, 0.2).
Try this example in the calculator
Center: a 95% interval
For a standard normal distribution, center mode with p = 0.95 gives z ≈ ±1.95996, the familiar ±1.96. For mean 100 and standard deviation 15, the middle 95% of values fall between about 70.60 and 129.40.
Try this example in the calculator
Two tails: critical values for α = 0.05
For a two-tailed test at α = 0.05, enter 0.05 in two-tails mode. The 0.05 splits into 0.025 per tail, and the critical values are z ≈ −1.95996 and z ≈ 1.95996.
Try this example in the calculator
Want more practice? See more inverse normal examples.
InvNorm vs NormalCDF: which one do you need?
| normalCdf | invNorm | |
|---|---|---|
| You enter | Values (lower, upper, μ, σ) | An area (probability), μ, σ |
| You get | An area (probability) | A value (z-score or x-value) |
| Example | normalCdf(−1E99, 1.96) ≈ 0.975 | invNorm(0.975) ≈ 1.96 |
| Question sounds like | "What proportion scores below 650?" | "What score is at the 90th percentile?" |
Rule of thumb: if your question gives a percentage and asks for a value, use invNorm. If it gives a value and asks for a percentage, use normalCdf and our normal CDF calculator. Full comparison: invNorm vs normalCdf.
InvNorm on your calculator or software
TI-84 Plus and TI-83
Press 2nd → VARS (DISTR) → 3:invNorm(, then enter invNorm(area, μ, σ). The area is the left-tail area unless your OS offers the tail option. For a right-tail 0.05, enter 0.95. Full TI-84 guide →
Casio
On the fx-9750GIII, fx-9860 and fx-CG50: STAT → DIST → NORM → InvN, then choose Left, Right or Central and enter the area, σ and μ. Menu labels vary slightly by model. Full Casio guide →
Excel and Google Sheets (NORM.INV, NORM.S.INV, NORMSINV)
Standard normal: =NORM.S.INV(p) (older name =NORMSINV(p)). Custom: =NORM.INV(p, μ, σ) (older =NORMINV). Google Sheets accepts both names. Full Excel guide →
R and Python (qnorm, scipy.stats.norm.ppf)
R: qnorm(p) or qnorm(p, mean = μ, sd = σ). Python: from scipy.stats import norm; norm.ppf(p, loc=μ, scale=σ). Both use the left-tail area, like the TI-84.
Common z critical values
| Confidence level | α (significance) | α/2 | Critical z (±) |
|---|---|---|---|
| 80% | 0.20 | 0.10 | ±1.28155 |
| 90% | 0.10 | 0.05 | ±1.64485 |
| 95% | 0.05 | 0.025 | ±1.95996 |
| 98% | 0.02 | 0.01 | ±2.32635 |
| 99% | 0.01 | 0.005 | ±2.57583 |
Need other levels or t and chi-square values? Use the critical value calculator, or build an interval with the confidence interval calculator.
Inverse normal formula and how accurate this calculator is
The formula: z = Φ⁻¹(p), x = μ + zσ
For the standard normal distribution, invNorm returns z = Φ−1(p), where Φ is the normal cumulative distribution function (CDF). For a normal distribution with mean μ and standard deviation σ, the cutoff is x = μ + z·σ. Φ−1 has no closed form, so every calculator uses a numerical method.
Method and verification
This calculator starts from Peter Acklam's rational approximation of Φ−1 and refines it with one Halley step against a double-precision normal CDF. Everything runs in your browser and nothing is sent to a server. Results agree with R's qnorm to at least 12 decimal places:
| p | This calculator | R qnorm(p) |
|---|---|---|
| 0.90 | 1.2815515655 | 1.2815515655 |
| 0.95 | 1.6448536270 | 1.6448536270 |
| 0.975 | 1.9599639845 | 1.9599639845 |
| 0.99 | 2.3263478740 | 2.3263478740 |
To 16 significant digits, Φ−1(0.95) = 1.644853626951473. Excel's NORM.S.INV and the TI-84's invNorm agree to the 5 decimals shown in the tables above. See the methodology page for the algorithm and references.
InvNorm FAQs
invNorm is undefined at exactly 0 or 1 because the normal curve never reaches its ends: the cutoff would be −∞ or +∞. A TI-84 returns a domain error, and this calculator asks for a value strictly between 0 and 1.
Yes. 0.95, .95, 95 and 95% are all read as a probability of 0.95. A TI-84 only accepts the decimal form, so type 0.95 there.
Follow your course. Many textbooks round z-scores to 2 decimals (the precision of a z table) and x-values to 2–4 decimals. A TI-84 displays up to 10 significant digits. Use the Decimal places menu to match what your teacher expects.
A left-tail probability below 0.5 means the cutoff sits below the mean, so z is negative. For example, invNorm(0.20) ≈ −0.84162. If you expected a positive value, check whether your question describes the area to the right; if so, use right-tail mode or enter 1 − p.
Tick "Use variance instead of standard deviation" in Custom Normal mode and enter the variance; the calculator takes the square root for you. On a TI-84, enter σ = √variance.
A printed z table only lists probabilities for z to 2 decimals, so a reverse lookup picks the nearest entry. This calculator solves for z directly, so it is more precise. For example, a table gives 1.645 for 0.95, while the exact value is 1.64485.
Related statistics calculators
Normal CDF Calculator
Calculate cumulative probabilities for normal distributions.
Z Score Calculator
Convert raw scores to Z scores and find probabilities.
Z Score to Percentile
Convert a Z score to its corresponding percentile.
Percentile to Z Score
Find the Z score for a given percentile.
Critical Value Calculator
Find critical values for confidence levels and hypothesis tests.
Confidence Interval Calculator
Calculate confidence intervals for population means.
Normal Distribution Calculator
Compute probabilities and values for any normal distribution.
Standard Normal Calculator
Work with the standard normal distribution (μ=0, σ=1).
Five Number Summary & IQR Calculator
Find min, Q1, median, Q3, max, IQR, and outliers.
All calculations run in your browser; no data is sent to a server. See our privacy policy.