P-Value Calculator · Practical guide
One-Tailed vs. Two-Tailed P Value
Same statistic, different question — and the choice has to be made before you look at the data.
At a glance
The Difference in One Line
A one-tailed p-value measures the chance of an effect in one specified direction; a two-tailed p-value measures it in either direction. For symmetric distributions like Z and T, the two-tailed value is exactly double the one-tailed value.
Interactive tool
P-Value Calculator
Enter your own values, then compare the result with the examples below.
P-Value Results
Same Statistic, Both Answers
For z = 1.96, the one-tailed p is 0.025 and the two-tailed p is 0.050. The same statistic can therefore be "significant" one way and not the other — which is exactly why the choice of tails must be made before looking at the data.
When Each Is Justified
- Two-tailed (the default): you want to detect a difference in either direction — a new drug could help or harm. Most published research uses two-tailed tests.
- One-tailed (the exception): your hypothesis specifies a direction in advance and the opposite outcome is meaningless — for example, testing only whether a process exceeds a safety threshold.
The Temptation to Avoid
Switching from two-tailed to one-tailed after seeing a borderline result (p = 0.06 quietly becoming 0.03) is a form of p-hacking. Pre-register your test type, then report the p-value that matches it. The calculator above shows both so you can verify, not shop.
Frequently Asked Questions
Is a one-tailed test ever more powerful?
Yes — concentrating all the rejection region in one tail makes it easier to reach significance in that direction. That power is only legitimate when the directional hypothesis was fixed in advance.
What is the two-tailed p if my one-tailed p is 0.04?
For symmetric distributions like Z and T, simply double it: 0.08. The calculator displays both so you never have to convert by hand.
Does the doubling rule apply to chi-square tests?
No. The chi-square distribution is asymmetric and defined only for non-negative values, so this calculator reports the right-tail probability and caps the two-tailed value at twice that, up to 1.
Which tail should I report in a paper?
Report whichever test type your methods section pre-specified — two-tailed in most fields. Report the exact p-value (p = 0.032) rather than just below 0.05.