Skip to content
Statculator
P-Values & Critical Values

Z Critical Value Calculator

Free z critical value calculator and table. Get z* for any α or confidence level (90%, 95%, 99%), one- or two-tailed, with steps and a shaded normal curve.

Free, no sign-up Runs in your browserBy Hira Noor, BS MathematicsUpdated

Try
Enter the

Total probability of a Type I error, for example 0.05.

Tail

Two-tailed for H₁: “not equal to” or a two-sided confidence interval; one tail for a directional test or a one-sided bound.

Results

Critical value z* (two-tailed)
±1.96
z* = Φ⁻¹(0.975)
Area in each tail
0.025
Significance level, α
0.05
Confidence level, 1 − α
95%
Cumulative area, Φ(z*)
0.975
Central area between −z* and z*
95%
One-tailed z* at the same α
1.6449
Decision rule
Reject H₀ if |z| ≥ 1.96

Interpretation: In a two-tailed z-test at α = 0.05, reject H₀ when |z| ≥ 1.96. The same multiplier gives a 95% confidence interval: estimate ± 1.96 standard errors.

normal distribution curve with the shaded probability region, markers at -1.96, z* = 1.96-1.96z* = 1.96-4-3-2-101234Standard normal distribution, Z ~ N(0, 1)
Shaded: rejection region with total area α = 0.05 (0.025 in each tail). Dashed lines: critical values.

z critical values for common confidence levels

Confidence levelαTwo-tailed z*One-tailed z*
80%0.21.28160.8416
90%0.11.64491.2816
95%0.051.961.6449
98%0.022.32632.0537
99%0.012.57582.3263
99.9%0.0013.29053.0902

Two-tailed z* = Φ⁻¹(1 − α/2), used for two-sided tests and intervals. One-tailed z* = Φ⁻¹(1 − α), used for directional tests and one-sided bounds (negate it for a left-tailed test).

Step-by-step solution

  1. 1

    Decide where α goes

    A two-tailed test (or two-sided interval) splits α equally between the two tails.

    α/2 = 0.05 / 2 = 0.025 in each tail
  2. 2

    Find the cumulative area to the left of the critical value

    P(Z ≤ z*) = 1 − α/2 = 0.975
  3. 3

    Invert the standard normal CDF

    Φ⁻¹ is the inverse normal function (invNorm on a TI-84, NORM.S.INV in Excel, qnorm in R).

    z* = Φ⁻¹(0.975) = 1.96
    By symmetry the lower critical value is −1.96, so z* = ±1.96
  4. 4

    Use the critical value

    Reject H₀ if |z| ≥ 1.96
    95% confidence interval: estimate ± 1.96 × SE

What is a Z Critical Value Calculator?

A z critical value calculator finds z*, the critical z value that leaves a chosen area α in the tail or tails of the standard normal curve. For a two-tailed test or a two-sided confidence interval, z* = Φ⁻¹(1 − α/2), so 95% confidence gives ±1.96 and 99% gives ±2.576. For a one-tailed test, z* = Φ⁻¹(1 − α), so α = 0.05 gives 1.645.

What the Z Critical Value Calculator does

Enter either a significance level α or a confidence level, choose the tail, and the calculator returns z* with the rejection region shaded on the standard normal curve. It also shows the tail area, the cumulative probability it inverted, the matching one- or two-tailed value, and z* for 80%, 90%, 95%, 98%, 99% and 99.9% confidence. It works as a critical z value calculator for any α, not only the levels printed in textbooks.

The same number goes by several names: the critical z value, the critical z-score, the z-score critical value, z star, z_c (zc), z_α or z_α/2. Textbook exercises such as “determine z_α for the following values of α”, “find the value of z subscript alpha” or “find the indicated critical z value” all ask for it. Choose the right tail for z_α and the two-tailed option for z_α/2.

z* is what you compare a z statistic with in the critical-value approach to hypothesis testing, and it is the multiplier in every z-based confidence interval and margin of error: estimate ± z* × SE. It applies whenever the sampling distribution is normal, including z-tests for means, proportion tests, proportion intervals and margin of error calculations.

How to use the Z Critical Value Calculator

  1. Choose whether to enter a Significance level α or a Confidence level.
  2. Type the value, for example 0.05 for α or 95 for a 95% confidence level.
  3. Pick the Tail: Two-tailed for “not equal to” tests and two-sided intervals, Right or Left for directional tests and one-sided bounds.
  4. Read z* and the shaded rejection region. The step-by-step solution shows how α becomes a cumulative probability and then a z-value.

What is a z critical value?

A z critical value is the point on the standard normal distribution, N(0, 1), that cuts off a tail area equal to the significance level. Values of the test statistic beyond it fall in the rejection region, so z* is the boundary between “reject H₀” and “do not reject H₀” in a z-test. In a confidence interval, the area between −z* and +z* equals the confidence level, so z* sets how many standard errors wide the interval is.

How to find the z critical value

To calculate a critical z value by hand you need the significance level α and the tail of the test. Then read z* from a z table or compute the normal quantile.

  1. Get α. Use the stated significance level, or for a confidence level C use α = 1 − C (95% → 0.05).
  2. Pick the tail. Two-tailed puts α/2 in each tail; one-tailed puts all of α in one tail.
  3. Find the cumulative area to the left of z*. Two-tailed: 1 − α/2. Right-tailed: 1 − α. Left-tailed: α.
  4. Invert the normal CDF. Find the z with that area to its left, using the z critical value table below, a z table read in reverse, or Φ⁻¹ in software. Example: α = 0.05 two-tailed gives a cumulative area of 0.975 and z* = ±1.96.

One-tailed or two-tailed: which z critical value do you need?

Choose the tail from the alternative hypothesis, before you look at the data:

  • Two-tailed (H₁: μ ≠ μ₀): use ±z* with α/2 in each tail. Reject when |z| ≥ z*. Use the two-tailed value for every two-sided confidence interval too.
  • Right-tailed (H₁: μ > μ₀): use +z* with all of α in the upper tail. Reject when z ≥ z*.
  • Left-tailed (H₁: μ < μ₀): use −z* with all of α in the lower tail. Reject when z ≤ −z*.
  • At α = 0.05 the one-tailed value, 1.645, is smaller than the two-tailed 1.96, so decide the tail from the research question, not to get an easier cut-off.
  • Not a z problem? Use the t critical value when σ is estimated from a sample, the chi-square critical value for counts and variances, or the general Critical Value Calculator to switch between distributions.

Z critical value table

This table of critical z values covers the significance and confidence levels used most often. The two-tailed column is z_α/2, for two-sided tests and intervals. The one-tailed column is z_α, the right-tailed value; put a minus sign in front of it for a left-tailed test. For any other α, use the calculator above.

Critical values of z by significance level α
αConfidence level (1 − α)Two-tailed z_α/2One-tailed z_α
0.5050%±0.67450.0000
0.2080%±1.28160.8416
0.1585%±1.43951.0364
0.1090%±1.64491.2816
0.0892%±1.75071.4051
0.0595%±1.96001.6449
0.0496%±2.05371.7507
0.02597.5%±2.24141.9600
0.0298%±2.32632.0537
0.0199%±2.57582.3263
0.00599.5%±2.80702.5758
0.00299.8%±3.09022.8782
0.00199.9%±3.29053.0902
0.000199.99%±3.89063.7190

Z critical value formula

The critical z value formula takes a quantile of the standard normal distribution, found with the inverse CDF Φ⁻¹. There is no closed form for Φ⁻¹, so tables and software compute it numerically.

Two-tailed (two-sided interval)
z* = Φ⁻¹(1 − α/2) → reject if |z| ≥ z*
Right-tailed
z* = Φ⁻¹(1 − α) → reject if z ≥ z*
Left-tailed
z* = Φ⁻¹(α) = −Φ⁻¹(1 − α) → reject if z ≤ z*
From a confidence level C
α = 1 − C
Formula symbols
SymbolMeaning
z*critical value (also written z_α, z_α/2 or z_c)
αsignificance level, the total area of the rejection region
Cconfidence level, for example 0.95
Φ⁻¹inverse of the standard normal CDF (the quantile function)

The subscript notation varies between textbooks. z_α/2 and z_0.025 both mean the value with an area of 0.025 to its right, which is 1.96.

Sources: NIST/SEMATECH e-Handbook, 1.3.6.6.1 Normal distribution; NIST/SEMATECH e-Handbook, 1.3.6.7.1 Cumulative distribution function of the standard normal distribution; Microsoft Support, NORM.S.INV function. Values on this page were checked against SciPy’s norm.ppf.

Worked example: two-sided confidence interval

z* for a 92% confidence interval

Problem. A survey team wants a 92% two-sided confidence interval for a proportion. What z* should they use?

  1. 01Convert to α: α = 1 − 0.92 = 0.08.
  2. 02Split α across both tails: α/2 = 0.04 in each tail.
  3. 03Cumulative area to the left of z*: 1 − 0.04 = 0.96.
  4. 04Invert the normal CDF: z* = Φ⁻¹(0.96) = 1.7507.

Answer. Use z* ≈ 1.7507, so the interval is p̂ ± 1.7507 × √(p̂(1 − p̂)/n). It sits between the familiar 90% value (1.6449) and the 95% value (1.96), as it should.

Worked example: determine z_α for the following values of α

A common textbook exercise asks for z_α, the value with area α to its right, for α = 0.0055, 0.09 and 0.663. Set the calculator to Right-tailed and enter each α, or work it out by hand: the area to the left of z_α is 1 − α.

  1. α = 0.0055: area to the left = 0.9945, so z_0.0055 = Φ⁻¹(0.9945) = 2.54.
  2. α = 0.09: area to the left = 0.91, so z_0.09 = Φ⁻¹(0.91) = 1.34.
  3. α = 0.663: area to the left = 0.337, so z_0.663 = Φ⁻¹(0.337) = −0.42. When α is above 0.5, z_α is negative, because more than half the curve lies to its right.

Worked example: right-tailed z-test for a proportion

A website converts 40% of visitors. After a redesign, 98 of 200 visitors convert (p̂ = 0.49). Test H₀: p = 0.40 against H₁: p > 0.40 at α = 0.01.

  1. Tail and critical value: the test is right-tailed, so all of α = 0.01 goes in the upper tail. z* = Φ⁻¹(0.99) = 2.3263.
  2. Standard error under H₀: √(0.40 × 0.60 / 200) = 0.03464.
  3. Test statistic: z = (0.49 − 0.40) / 0.03464 = 2.60.
  4. Decision: 2.60 ≥ 2.3263, so z lands in the rejection region. Reject H₀ (p ≈ 0.0047). The One Proportion Z-Test Calculator runs the whole test.

Z critical values in Excel, R, Python and on a TI-84

Every package has an inverse-normal function that takes a cumulative (left-tail) probability. Feed it 1 − α/2 for a two-tailed value or 1 − α for a one-tailed value. The examples use α = 0.05, which return 1.96 two-tailed and 1.645 one-tailed.

ToolTwo-tailed z*One-tailed z* (right)
Excel / Google Sheets=NORM.S.INV(0.975)=NORM.S.INV(0.95)
TI-83 / TI-84invNorm(0.975)invNorm(0.95)
Rqnorm(0.975)qnorm(0.95)
Python (SciPy)norm.ppf(0.975)norm.ppf(0.95)

Critical values and p-values give the same decision

Comparing |z| with z* and comparing the p-value with α are two views of the same rule. If |z| = 2.10 in a two-tailed test at α = 0.05, it exceeds z* = 1.96, and its p-value of 0.0357 is below 0.05. The critical value is handy when you need the cut-off before seeing data, for example to plan a sample size; the p-value tells you how far past the cut-off you are.

The conversion also runs backwards. To turn a p-value into a z-score, use z = Φ⁻¹(1 − p) for a one-tailed p-value and z = Φ⁻¹(1 − p/2) for a two-tailed one. A two-tailed p of 0.0357 gives |z| = 2.10.

Common use cases

  • Building z confidence intervals for means with known σ and for proportions.
  • Setting rejection regions for z-tests before the data are collected.
  • Computing a margin of error or a required sample size, where z* is a direct input.
  • Reading off z_α values for power and sample-size formulas.
  • Checking values from a printed z table or a textbook appendix.

Assumptions and limitations

  • z* assumes the statistic is exactly or approximately normal. When σ is estimated from a small sample, use the t critical value instead; it is larger and gives the correct coverage.
  • For proportions, a z-based interval is unreliable when n·p̂ or n·(1 − p̂) is small; the Wilson interval in the Proportion Confidence Interval Calculator performs better.
  • One-tailed critical values must be chosen before seeing the data. Switching from 1.96 to 1.645 after seeing which way the data point doubles the real Type I error rate, from 5% to 10%.

Frequently asked questions

What is a z critical value?

A z critical value is the point on the standard normal curve that leaves the significance level α in the tail or tails. A z statistic beyond it falls in the rejection region, and in a confidence interval it is the number of standard errors on each side of the estimate, for example 1.96 at 95% confidence.

How do you find the z critical value?

Find the z-score that leaves the significance level in the tail or tails. For a two-tailed test put α/2 in each tail, so α = 0.05 gives z* = ±1.9600; for a one-tailed test put all of α in one tail, which gives 1.6449 at α = 0.05. To start from a confidence level, use α = 1 − confidence.

What is the z critical value for a 95% confidence interval?

For a two-sided 95% confidence interval, z* = 1.96 (more precisely 1.959964). It leaves 2.5% in each tail of the standard normal curve. For a one-sided 95% bound the value is 1.6449.

What is the z critical value for 99% confidence?

For a two-sided 99% confidence interval, z* = 2.5758, usually rounded to 2.576. It leaves 0.5% in each tail. A one-sided 99% bound, or a one-tailed test at α = 0.01, uses 2.3263.

What is the critical value of z at α = 0.05?

It depends on the tail. A two-tailed test at α = 0.05 uses ±1.96, a right-tailed test uses 1.6449, and a left-tailed test uses −1.6449. In each case the shaded rejection region has a total area of 0.05.

What z* do I use for 90% confidence?

Use z* = 1.6449 for a two-sided 90% confidence interval. It leaves 5% in each tail. Higher confidence requires a larger z*, which makes the interval wider: 1.96 at 95% and 2.5758 at 99%.

What does z_α mean?

z_α (z subscript alpha) is the z-value with an area of α to its right on the standard normal curve, so it is the right-tailed critical value. For example z_0.05 = 1.6449 and z_0.025 = 1.96. z_α/2 is the two-tailed critical value for significance level α.

How do you calculate a z-score from a p-value?

Invert the normal CDF. For a one-tailed (right-tail) p-value, z = Φ⁻¹(1 − p); for a two-tailed p-value, |z| = Φ⁻¹(1 − p/2). For example, a two-tailed p of 0.05 gives |z| = 1.96, and a one-tailed p of 0.03 gives z = 1.88.

Why is a left-tailed critical value negative?

A left-tailed test rejects H₀ for unusually small z values, so the cut-off sits below 0. Because the normal curve is symmetric, the left-tailed z* is just the negative of the right-tailed value, for example −1.6449 at α = 0.05.

When should I use a t critical value instead of z?

Use t* when the population standard deviation is unknown and estimated from the sample, which is the usual situation for means. t* is larger than z* for small samples, for example 2.086 versus 1.96 at 95% with 20 df, and approaches z* as df grows.

Cite this page

Statculator. (2026, November 14). Z Critical Value Calculator. Retrieved November 14, 2026, from https://statculator.com/p-values-critical-values/z-critical-value-calculator/