🧪 P-Value Calculator
P Value Calculator
The p-value is the probability of seeing a result at least this extreme if nothing were really going on. From your z-score, this calculator finds the two-tailed area under the bell curve — the standard yardstick of statistical significance.
The usual bar is p < 0.05: below it, researchers call the result statistically significant; above it, the evidence is too weak to rule out chance. The calculator labels your result against the 1%, 5% and 10% levels automatically.
How to use this calculator
Type your z-score and press Calculate for the two-tailed p-value and its significance reading.
- Type the z-score from your test.
- Press Calculate.
- Read the two-tailed p-value.
- Check the interpretation line for the 1, 5 and 10 percent significance levels.
Worked example: z = 1.96, every tail at once
One z-score, five honest questions — the result shows all five, because different studies ask different ones:
| Question | Value |
|---|---|
| Left tail: P(Z < 1.96) | 0.9750 |
| Right tail: P(Z > 1.96) | 0.0250 |
| Two tails: P(|Z| > 1.96) | ≈ 0.0500 |
| Between −1.96 and 1.96 | 0.9500 |
| From the centre to 1.96 | 0.4750 |
This is why 1.96 is famous: it is the z that puts exactly 2.5% in each tail and 95% in the middle — the boundary of the standard 5% test.
What does this p-value calculator compute?
It converts a test statistic into tail probabilities and a verdict at the significance level α you enter, using the distribution you choose: z, t, chi-square (χ²) or F. Z mode reports the left tail, right tail, two-tailed probability, probability between −z and z, and area from the centre to z, and its verdict uses the two-tailed p-value. T mode reports the left, right and two-tailed probabilities, with degrees of freedom, and its verdict also uses the two-tailed p-value.
Chi-square and F modes report the left and right tails, and their verdicts use the right-tail p-value; chi-square needs degrees of freedom, while F needs numerator and denominator degrees of freedom (df1 and df2). Every verdict compares the unrounded p-value with α, and the verdict line prints p to six decimal places so the displayed value and the verdict agree.
What a p-value actually is
Suppose the dull explanation is true — no effect, no difference, nothing going on (the null hypothesis). The p-value is the probability that pure chance would still produce a result at least as extreme as the one you saw. Small p: chance is a poor explanation, so the interesting one earns a hearing. Large p: chance explains it comfortably. It is a measure of surprise, not a probability that your hypothesis is true — that misreading is the most common error in all of applied statistics.
Worked example: a t-score instead
A t of 2.00 with 15 degrees of freedom gives a two-tailed p ≈ 0.0639 — larger than the p for the same-sized z (which would be about 0.0455), because the t-distribution has heavier tails: with little data, extreme results happen by chance more often. And a chi-square of 3.84 with 1 degree of freedom gives p ≈ 0.0500 — 3.84 is to chi-square what 1.96 is to z, and it is no coincidence: 1.96² = 3.84.
Reading the verdict
Compare p with your chosen significance level α (0.05 by default, set it before you look, not after): p < α → statistically significant at that level; the data would be surprising if nothing were going on. p ≥ α → not significant; chance remains a comfortable explanation. “Not significant” never proves there is no effect — it proves this dataset cannot convict one. A bigger sample is the usual remedy, which is why our t-test example (t = 2.00, p ≈ 0.064) reads as “promising, get more data” rather than “nothing there”.
Frequently asked questions
What is the p-value for z = 1.96?
It depends on the tail you mean: right-tailed 0.0250, left-tailed 0.9750, and two-tailed ≈ 0.0500 — which is exactly why 1.96 marks the 5% boundary. The calculator prints all five readings (including 0.95 between −1.96 and 1.96) so the phrasing of your hypothesis chooses, not the tool.
Why does t = 2.00 give a bigger p-value than z = 2.00?
Because the t-distribution has heavier tails — with a small sample (here df 15), chance produces extremes more often, so the same score is less surprising: p ≈ 0.0639 instead of about 0.0455. As degrees of freedom grow, the t-values converge to the z-values.
What does a chi-square p-value need that a z does not?
A distribution choice and degrees of freedom. Chi-square and F tests arise from variance and table problems rather than means, so the calculator asks for the distribution (z, t, χ² or F) and the df before computing. For χ² = 3.84 at df 1, p ≈ 0.0500.
Is p = 0.064 “nearly significant”? Should I report it as a trend?
Report the number itself, plainly: p = 0.064. “Nearly significant” as a rhetorical upgrade is how weak results get oversold. The honest sentence is that the data are suggestive but do not meet the 5% level — and the constructive next step is a larger sample, not a softer adjective.
Does a small p-value prove my hypothesis?
No. A small p says the data would be surprising if the null hypothesis were true. It says nothing about whether your specific alternative is the right explanation, how large the effect is, or whether the study was designed well. Significance is the start of an argument, not the end of one.







































