P-value Calculator

Convert a z or t statistic into a tail probability, or work backwards from a probability to a statistic. Compare regions and understand what the result means.

Two-tailed p-value0.0499957903

-1000000000000 – 1000000000000

0.000001 – 0.5

Two-tailed p-value

0.0499957903

  • Left-tail probability0.9750021049
  • Right-tail probability0.0249978952
  • Two-tailed p-value0.0499957903
  • Area from zero to |score|0.4750021049
  • Area between ±|score|0.9500042097

A p-value is not the probability that the null hypothesis is true, and does not measure effect size. Choose the test and tail before looking at results. Two-tail and central inverse calculations return the nonnegative magnitude (both signs have the same area). Extremely small probabilities can underflow to zero at machine precision.

How this was calculated

Reference: standard normal; statistic = 1.96.

Left = F(score); right = 1 − F(score); two tails = 2 × min(left, right). Central area = 1 − two tails; area from zero = central area ÷ 2.

p = 0.0499957903 ≤ α = 0.05: reject the null at this chosen level, if the test assumptions hold.

Choose the reference distribution and tail

A z statistic is compared with the standard normal distribution. A t statistic is compared with Student’s t distribution using the degrees of freedom of the underlying test. This tool converts an already computed statistic; it does not decide which test fits your data.

If F is the cumulative distribution function, left-tail probability is F(s), right-tail probability is 1 − F(s), and the two-tailed value for these symmetric distributions is twice the smaller tail. Choose the alternative hypothesis before observing the result.

A worked normal example

For z = 1.96, the left-tail area is about 0.9750021, the right-tail area is about 0.0249979, and the two-tailed p-value is about 0.0499958. A two-tailed comparison with a preselected 0.05 significance level just meets that threshold.

In reverse mode, a normal two-tailed probability of 0.05 corresponds to a magnitude of about 1.959963985. Both positive and negative statistics of that magnitude have the same two-tailed area. Central-area modes show probability coverage, not hypothesis-test p-values.

Interpreting the output

A p-value describes how extreme a result would be under the null model and the test’s assumptions. It is not the chance that the null hypothesis is true, the probability an effect is important, or a substitute for effect size and study design.

The implementation computes small tails directly to avoid subtraction loss and numerically inverts the selected region. Inverse probabilities are supported from 10^-12 to 1 − 10^-12; Student t supports 1–10,000 whole degrees of freedom. Extreme forward tails can fall below floating-point precision.

How to use the P-value Calculator

Convert a z or t statistic into a tail probability, or work backwards from a probability to a statistic. Compare regions and understand what the result means.

  1. Choose the distribution and direction

    Select normal or Student t, the probability region, and whether to solve for probability or statistic.

  2. Enter your statistic or probability

    Supply degrees of freedom for Student t and your chosen significance level for tail tests.

  3. Interpret the selected area

    Review the left, right and central regions. Use the stated assumptions when interpreting a p-value.

References

Frequently asked questions

Does p above 0.05 prove there is no effect?

No. Not rejecting a null hypothesis does not establish that it is true. Consider precision, sample size, test assumptions and the size of any estimated effect.

Why does an inverse two-tailed calculation return a positive value?

It returns the nonnegative magnitude. Both signs have the same two-tailed probability; a directional left- or right-tail calculation retains the appropriate sign.

What are degrees of freedom?

They are a parameter of the t reference distribution determined by the statistical test. For a basic one-sample t test they are often sample size minus one, but other tests use different formulas.

Last updated . Results are estimates for informational purposes only.