P Value Visualizer
P-value visualization
Test Statistics
Decision
Click any scenario to load it into the visualizer. Each example uses realistic data from education, medicine, business, and research.
Two-tailed
One-tailed (right)
Key Difference
The one-tailed p-value is exactly half the two-tailed p-value for the same test statistic and symmetric distribution.
P-Value Examples by Field
Real scenarios across medicine, education, business, and psychology — click any card to load it
What Is a P-Value?
A p-value is the probability of obtaining a test statistic at least as extreme as the one observed in your sample, assuming the null hypothesis is true. It is a conditional probability — it tells you how surprising your data would be if there were actually no effect. The p-value does not tell you the probability that your hypothesis is correct, nor does it measure the size or importance of an effect.
Formally: given a test statistic T computed from your data, the p-value is P(|T| ≥ |t_observed| | H₀ is true). When this probability is small (below your chosen α), you have evidence against the null hypothesis. When it is large, your data are consistent with H₀ — but that does not prove H₀ is true.
How the Visualization Works
The visualizer draws the sampling distribution of your test statistic (normal or t). The shaded region in red represents the p-value: the area under the curve that is as extreme or more extreme than your observed test statistic. The vertical dashed line shows your test statistic, and the dotted lines show the critical values determined by your alpha level. When the test statistic falls inside the rejection region (past the critical value), the result is statistically significant at your chosen α.
What Does the Shaded Area Represent?
In the graph, the shaded area is the p-value itself expressed as a proportion of the total area under the curve. Because the curve represents a probability distribution, its total area equals 1. The red shaded tail (or tails, in a two-tailed test) represents all the outcomes that would be at least as extreme as what you observed. If you ran the experiment many times under H₀, this shaded proportion is how often you would see results this extreme just by chance.
One-Tailed vs Two-Tailed Tests
| Feature | Two-Tailed | One-Tailed (Right) | One-Tailed (Left) |
|---|---|---|---|
| Hypothesis | H₁: μ ≠ μ₀ | H₁: μ > μ₀ | H₁: μ < μ₀ |
| Shaded region | Both tails | Right tail only | Left tail only |
| p-value calculation | 2 × P(Z ≥ |z|) | P(Z ≥ z) | P(Z ≤ z) |
| Critical value (α=0.05) | ±1.96 | 1.645 | −1.645 |
| When to use | No prior direction assumed | Expect positive effect | Expect negative effect |
| Conservative? | Yes (standard default) | No — easier to reject H₀ | No — easier to reject H₀ |
How to Interpret P-Values: A Decision Guide
Z Distribution vs t Distribution
Use the z distribution when the population standard deviation is known, or when your sample size is large (n > 30 is a common rule of thumb). Use the t distribution when the population standard deviation is unknown and you are estimating it from your sample. The t distribution has heavier tails than the normal distribution, which accounts for the extra uncertainty from estimating σ. As degrees of freedom increase, the t distribution approaches the standard normal.
Common Misconceptions About P-Values
- Misconception: "p = 0.04 means there is a 4% chance the null hypothesis is true." The p-value is not the probability that H₀ is true. It is the probability of your data (or more extreme data) given that H₀ is true.
- Misconception: "A non-significant result proves there is no effect." Failing to reject H₀ only means you lack enough evidence — not that the effect is zero. Absence of evidence is not evidence of absence.
- Misconception: "A smaller p-value means a bigger effect." p-values are sensitive to sample size. A tiny effect can produce p < 0.001 with a very large sample.
- Misconception: "p = 0.049 and p = 0.051 represent fundamentally different results." These two values are nearly identical statistically. The 0.05 threshold is a convention, not a natural law.
- Misconception: "A p-value can be greater than 1." It cannot — p-values are probabilities bounded between 0 and 1.
The SHADE Framework: Reading P-Value Graphs Step by Step
Entity & Glossary Reference
| Term | Symbol | Definition |
|---|---|---|
| P-value | p | Probability of observing a test statistic as extreme or more extreme than calculated, given H₀ is true. Range: [0, 1]. |
| Alpha level | α | The significance threshold set before the study. Typical values: 0.10, 0.05, 0.01. Represents acceptable Type I error rate. |
| Null hypothesis | H₀ | The default claim of no effect or no difference. The hypothesis you are testing against. |
| Alternative hypothesis | H₁ or Hₐ | The research hypothesis — the effect or difference you are trying to detect. |
| Test statistic | z or t | A standardized score computed from sample data, measuring how far the sample estimate is from the null hypothesis value. |
| Critical value | z* or t* | The boundary of the rejection region. Determined by α and the distribution. For α=0.05 two-tailed: z* = ±1.96. |
| Rejection region | — | The set of test statistic values that lead to rejecting H₀. Corresponds to the shaded tail area in the visualizer. |
| Type I error | α | Rejecting H₀ when it is actually true (false positive). Probability equals α. |
| Type II error | β | Failing to reject H₀ when it is actually false (false negative). |
| Statistical power | 1 − β | Probability of correctly rejecting a false H₀. Increases with sample size and effect size. |
| Confidence level | 1 − α | The proportion of confidence intervals that would contain the true parameter across repeated samples. |
| Effect size | d, η², r | A standardized measure of how large the effect is, independent of sample size. Cohen's d is common for means. |
Related Topics
Sources & further reading:
- Wasserstein, R. L., & Lazar, N. A. (2016). The ASA Statement on p-Values. The American Statistician, 70(2), 129–133. — American Statistical Association
- NIST Engineering Statistics Handbook — What are the steps for hypothesis testing?
- Khan Academy — P-values and significance tests
- Penn State STAT 415 — Introduction to Mathematical Statistics
- OpenStax Statistics — Introductory Statistics (free textbook)
Frequently Asked Questions
A p-value is the probability of observing a test statistic as extreme as, or more extreme than, the one calculated from your data, assuming the null hypothesis is true. It measures how surprising your sample result would be under H₀. A p-value of 0.03 means: if H₀ were true, you would only see data this extreme about 3% of the time by chance alone.
The curve drawn in the visualizer is the probability density function of your test statistic under H₀. Area under the curve equals probability. The shaded tail represents all outcomes as extreme or more extreme than your observed statistic — which is precisely what the p-value measures. A larger shaded area means a less surprising result; a smaller shaded area means your data would be unusual if H₀ were true.
A two-tailed test shades both tails of the distribution and tests whether the effect goes in either direction. A one-tailed test shades only one tail and tests for an effect in a specific direction. For a given test statistic, the one-tailed p-value is exactly half the two-tailed p-value in a symmetric distribution. Most scientific publications default to two-tailed tests because they are more conservative and do not require prior directional assumptions.
Use a z-test when the population standard deviation (σ) is known, or when your sample size is large enough (generally n > 30) that the sample SD reliably estimates σ. Use a t-test when σ is unknown and you are estimating it from your sample — which is the common situation in practice. The t-distribution has heavier tails to account for this extra uncertainty; as df increases, it converges to the standard normal. This visualizer supports both.
No. Rejecting H₀ only means your data are inconsistent with the null hypothesis at your chosen significance level. It does not prove H₁. You could still have a Type I error (falsely rejecting a true H₀), which happens with probability α. Statistical significance also does not mean practical significance — always report effect sizes and confidence intervals alongside p-values to give a complete picture of your results.
A smaller alpha (e.g. 0.01) sets a stricter threshold — the critical value moves further into the tails, so the rejection region becomes narrower. Your test statistic needs to be more extreme to achieve significance. A larger alpha (e.g. 0.10) makes it easier to reject H₀ but increases the risk of a Type I error. Use the alpha slider in the visualizer above to see exactly how the critical value lines shift as you adjust α.
No. A p-value is a probability, and probabilities are always between 0 and 1. A p-value of exactly 0 would mean the observed result is impossible under H₀; a p-value of 1 would mean every possible result is at least as extreme as what you observed. In practice, p-values reported as "0.000" in software output are not exactly zero — they are simply smaller than the software's display precision.
They are mathematically equivalent: if a 95% confidence interval does not include the null hypothesis value, then p < 0.05 for a two-tailed test. Both use α as their threshold. Confidence intervals are generally preferred in reporting because they convey both the direction and magnitude of an effect, not just a binary significant/non-significant verdict. See our Confidence Interval Visualizer for a side-by-side comparison.