Z Score to P Value Calculator
Enter a z-score from your hypothesis test and get every p-value at once: one-tailed left, one-tailed right, two-tailed, the matching percentile, and a plain-English significance check. Free — no signup.
Convert z-score to p-value
Type the z test statistic from your analysis — sign included. The calculator shows all tail options so you can pick the one your hypothesis needs.
The shaded tails show the two-tailed p-value: the area at least as extreme as your z-score, in either direction.
This z score to p value calculator runs entirely in your browser. Nothing is uploaded or stored.
What is a p-value?
Every hypothesis test ends with the same question: "Could this just be chance?" The p-value answers it with a number between 0 and 1. A p-value of 0.03 says: if there were really no effect at all, you would still see a result this extreme about 3% of the time just by luck. A p-value of 0.60 says the result is thoroughly ordinary under the null — nothing surprising happened. Two warnings belong with every p-value. First, it is not the probability that the null hypothesis is true, and it is not the probability that your result happened "by chance" in plain language. Second, a p-value says nothing about how big or important an effect is — a tiny effect can have a tiny p-value if the sample is huge. The p-value only measures how incompatible the data are with the null hypothesis.
How the z-to-p conversion works
When a test statistic follows (approximately) the standard normal distribution, converting it to a p-value is just measuring tail area under the bell curve. There are only three areas to choose from, and the calculator above reports all three every time:
- One-tailed, right: p = 1 − Φ(z) — the area above your z-score. Use it when your alternative hypothesis says the true value is greater than the null value.
- One-tailed, left: p = Φ(z) — the area below your z-score. Use it when the alternative says the true value is smaller.
- Two-tailed: p = 2 × (1 − Φ(|z|)) — the area beyond |z| in both directions. Use it when an effect in either direction would count against the null. This is the default in most published research.
That is the whole mechanic — there is no separate "p-value formula" to memorise, just the tail area that matches the hypothesis you stated before seeing the data. If you still need the z-score itself from raw data, start with the z score calculator; if you need the cutoff a test must beat rather than the p-value of a result, use the critical z score calculator.
One-tailed vs two-tailed p-values
Choosing the wrong tail halves or doubles your p-value — it leads the list of one-tail vs two-tail mistakes that catch people out.
The same z-score produces three different p-values, and choosing between them is the most consequential decision in the whole calculation. For z = 1.96:
| Test type | P-value | Significant at α = 0.05? | When it is the right choice |
|---|---|---|---|
| One-tailed (right) | 0.0250 | Yes | You predicted a higher value in advance |
| One-tailed (left) | 0.9750 | No | You predicted a lower value — but z is positive, so the data point the other way |
| Two-tailed | 0.0500 | Yes (exactly on the line) | An effect in either direction matters |
The rule is simple but strict: decide the tail before you run the test. A two-tailed test splits the "extremeness" across both ends of the curve, which is why its p-value is exactly double the one-tailed value for the same |z|. Choosing a one-tailed test after seeing which direction the data went quietly halves your p-value and invalidates the test — examiners and reviewers both watch for it. If the direction was genuinely fixed in advance (a preregistered prediction, a one-sided safety check), a one-tailed test is legitimate and slightly more powerful. Everything else should be two-tailed.
How to use this calculator
- Run your z-test first. You need a test statistic z — from a sample mean, a proportion, or any approximately normal statistic. No z yet? The formula page shows how to compute one, including the sample-mean version with standard error.
- Type the z-score into the box, sign included. Results update as you type.
- Read the p-value that matches your hypothesis — left, right, or two-tailed — not simply the smallest one on the screen.
- Compare with your α (usually 0.05). The calculator states the verdict for α = 0.05 and α = 0.01 in plain English.
- Report the exact p-value in your write-up (for example "z = 2.58, p = 0.0099, two-tailed"), not just "significant".
Worked examples
Example 1 — a result that clears both bars. A tutoring programme claims to raise scores. A one-sample z-test on the results gives z = 2.58. The two-tailed p-value is 2 × (1 − Φ(2.58)) = 2 × (1 − 0.9951) = 0.0099. That is below 0.01 as well as 0.05, so the result is statistically significant at both conventional levels: if the programme did nothing, a result this extreme would occur less than 1% of the time.
Example 2 — a result that clears neither. A/B test on a web page: the new design converts slightly better, z = 0.50. The two-tailed p-value is 2 × (1 − Φ(0.50)) = 2 × (1 − 0.6915) = 0.6171. Nowhere near significant — differences this size appear by chance more than 60% of the time. The correct conclusion is "no evidence of a difference", not "the designs are equal".
Example 3 — a negative z-score. A manufacturer checks whether a filler under-fills bottles; the test statistic is z = −1.50 with a left-tailed alternative ("less than the label weight"). The one-tailed p-value is Φ(−1.50) = 0.0668. Interesting, but above 0.05 — not significant at the 5% level. Note how the sign matters here: the same |z| = 1.50 in a right-tailed test would give p = 0.9332, because the data point in the opposite direction from that alternative.
P-value vs percentile — do not mix them up
Students constantly confuse these because both come from the same curve. A percentile describes a position in a distribution: z = 1.96 is the 97.5th percentile, meaning 97.5% of values fall at or below it. A p-value answers a testing question: how extreme is this result in the tail(s) my hypothesis cares about? For z = 1.96 the percentile is 97.5%, the right-tailed p-value is 0.025, and the two-tailed p-value is 0.050 — three different numbers with three different jobs. If a question asks "what percentage scored below this?", it wants the percentile (use the z score to percentile calculator). If it asks "is this result significant?", it wants a p-value.
Common mistakes
- Picking the tail after seeing the data. Decided in advance, or it doesn't count.
- Reporting the percentile as the p-value. "p = 0.975" for z = 1.96 is actually the percentile in disguise — see the section above.
- Doubling wrongly. The two-tailed p-value is 2 × the smaller tail area (equivalently 2 × (1 − Φ(|z|))), not 2 × the one-tailed p for the direction you ran.
- Reading p as the probability the null is true. The p-value assumes the null is true; it cannot then tell you the probability of that assumption.
- Treating 0.05 as a cliff edge. p = 0.049 and p = 0.051 are nearly identical evidence. Report the number.
Still getting comfortable with the basics? The z-score FAQs answer the questions students ask most, in plain language.
Frequently asked questions
1. What is a p-value?
A p-value is the probability of getting a result at least as extreme as the one you observed, assuming the null hypothesis is true. A small p-value (typically ≤ 0.05) means the observed result would be surprising if nothing were really going on, so researchers treat it as evidence against the null hypothesis. It is not the probability that the null hypothesis is true.
2. How do you convert a z-score to a p-value?
Find the area under the standard normal curve in the tail (or tails) beyond the z-score. For a right-tailed test the p-value is 1 − Φ(z); for a left-tailed test it is Φ(z); for a two-tailed test it is 2 × (1 − Φ(|z|)). For z = 1.96 the two-tailed p-value is 0.0500.
3. Should I use a one-tailed or two-tailed p-value?
Decide before you look at the data. Use a two-tailed test when an effect in either direction would matter — this is the default in most research. Use a one-tailed test only when you had a directional prediction in advance and an effect in the opposite direction would be treated as no effect at all.
4. Is a p-value the same as a percentile?
No. A percentile says what share of values falls at or below a score — for z = 1.96 that is 97.5%. A p-value measures how extreme the result is in the tail(s) relevant to your hypothesis: the right-tailed p-value for z = 1.96 is 0.025 and the two-tailed p-value is 0.05. Percentiles describe position; p-values test hypotheses.
5. What p-value counts as statistically significant?
By convention, p ≤ 0.05 is called statistically significant and p ≤ 0.01 highly significant, but the threshold (α) should be chosen before the test and suited to the field. A p-value of 0.049 and one of 0.051 are practically identical evidence — report the actual p-value, not just whether it crossed a line.