Inverse Z-Score Calculator

Input Type

Find z such that P(Z ≤ z) = p (e.g., percentile cutoff).

Solution

—
Share:

Worked Examples

Confidence Interval

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

The most cited critical value in statistics. Used for two-sided 95% confidence intervals and α = 0.05 two-tailed tests.

  1. Choose the Confidence level mode.
  2. Enter 95.
  3. The middle area equals 95%, so each tail equals 2.5%.
  4. The inverse normal CDF returns ±1.96.
  5. These are the boundaries of the standard 95% confidence interval.

±1.96 multiplied by the standard error gives the half-width of a 95% confidence interval.

Hypothesis Testing

What is the critical z for a one-tailed test at α = 0.05?

A common upper-tail rejection threshold for one-sided tests where you only care about deviations in one direction.

  1. Choose the Right-tail probability mode.
  2. Enter p = 0.05.
  3. Find the z-score with right-tail area equal to 0.05.
  4. The inverse normal CDF returns about 1.6449.
  5. Reject the null hypothesis if the test statistic exceeds 1.6449.

1.6449 is the standard one-tailed 5% critical z, distinct from the two-tailed value of 1.96.

Percentile

What z-score corresponds to the 90th percentile?

Use this when you have a percentile rank and need to translate it into a z-score for further calculation.

  1. Choose the Left-tail probability mode.
  2. Enter p = 0.90 (since the 90th percentile means 90% of values fall below it).
  3. The inverse normal CDF returns about 1.2816.
  4. So a z-score of 1.2816 marks the 90th percentile of the standard normal distribution.

Multiply this z-score by σ and add μ to find the equivalent raw value in any normal distribution.

Inverse Normal CDF

The inverse normal CDF — also called the probit or quantile function — returns the z-score that has a given cumulative probability under the standard normal distribution. Φ⁻¹(0.95) ≈ 1.6449 means 95% of standard normal values lie at or below z = 1.6449.

z = Φ⁻¹(p)

How It Works

An inverse z-score calculator runs the standard normal distribution in reverse. Instead of starting with a z-score and looking up its probability, you start with a probability, alpha level, or confidence percentage and find the z-score that produces it. This is the value statisticians call the critical value. The calculation uses the inverse normal CDF (the quantile or probit function), which is the mathematical inverse of the cumulative probability you would read off a z-table. Pick the mode that matches the question you are asking — left-tail probability, right-tail probability, two-tailed alpha, or confidence level — and the calculator returns the z-score that bounds the area you specified.

Example Problem

A researcher wants the critical z-score for a two-tailed test at α = 0.05. What is the cutoff value beyond which a result is considered statistically significant?

  1. Identify the test type: two-tailed at α = 0.05.
  2. Split alpha across the two tails: each tail gets α/2 = 0.025.
  3. Find the z-score whose right-tail area is 0.025 (or equivalently, whose left-tail area is 0.975).
  4. Apply the inverse normal CDF: z = Φ⁻¹(0.975) ≈ 1.9600.
  5. By symmetry, the critical values are ±1.96.
  6. Reject the null hypothesis if the test statistic falls below -1.96 or above +1.96.

This is why ±1.96 is the universally cited critical value for two-tailed 95% tests — it is the inverse z-score for α = 0.05 split across two symmetric tails.

Key Concepts

The inverse normal CDF answers the question: which z-score has this probability? It complements the forward normal CDF, which answers: what is the probability for this z-score? Together they form a two-way map between standardized scores and tail areas. Critical values are inverse z-scores commonly used as decision thresholds in hypothesis testing and confidence intervals. The most common ones are 1.6449 (one-tailed 95%), 1.96 (two-tailed 95%), 2.3263 (one-tailed 99%), and 2.5758 (two-tailed 99%).

Applications

  • Finding critical z-values for one-tailed and two-tailed hypothesis tests
  • Determining margin-of-error multipliers for confidence intervals
  • Setting cutoffs for percentile-based admissions or screening
  • Computing process control limits in quality assurance
  • Calibrating risk thresholds in finance, insurance, and reliability engineering
  • Building lookup tables for standardized scores in psychometrics

Common Mistakes

  • Forgetting to split alpha by 2 in two-tailed tests (using α instead of α/2)
  • Mixing up left-tail probability and right-tail probability — they are complements
  • Treating a confidence level (e.g., 95%) as a probability without dividing by 100
  • Reading a one-tailed critical z-value as if it were two-tailed
  • Using p ≥ 1 or p ≤ 0 — the inverse CDF is undefined at the endpoints
  • Confusing the inverse z-score (a critical value) with the test statistic itself

Frequently Asked Questions

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

For a two-sided 95% confidence interval, the critical z-score is ±1.96. For a one-sided 95% bound, it is approximately 1.6449.

What is the critical z-score for a 99% confidence interval?

For a two-sided 99% interval, the critical z-score is ±2.5758. For a one-sided 99% bound, it is about 2.3263.

What is the difference between a critical z-score and a test statistic?

A critical z-score is a fixed cutoff calculated from your alpha or confidence level. A test statistic is the z-score computed from your sample data. You compare the test statistic to the critical z-score to decide whether to reject the null hypothesis.

How is the inverse normal CDF calculated?

There is no closed-form formula. Most calculators use a high-accuracy rational approximation such as Acklam's algorithm. This calculator uses Acklam's approximation, accurate to about 1.15 × 10⁻⁹.

Why is α split in half for a two-tailed test?

A two-tailed test rejects the null hypothesis when the test statistic falls into either tail. To keep the total Type I error at α, half of α is allocated to the left tail and half to the right tail.

What probability inputs are valid?

For probability and alpha modes, the value must be strictly between 0 and 1. For confidence level, the value must be strictly between 0 and 100. The inverse CDF diverges to infinity at the boundaries.

What is the inverse z-score for the 90th percentile?

The 90th percentile corresponds to a left-tail probability of 0.90, which gives a critical z-score of approximately 1.2816.

How do I find the z-score for a one-tailed test at α = 0.01?

Use the right-tail mode with p = 0.01. The result is z ≈ 2.3263, the standard one-tailed 1% significance critical value.

Reference: The inverse normal CDF (probit) is computed using Peter Acklam's rational approximation, which has accuracy better than ~1.15 × 10⁻⁹ across the full unit interval.

Related Calculators

Related Sites