P-value to Z-score Calculator
What this P-value to Z-score tool does
A finished hypothesis test hands you a p-value, but significance tables, power planning, and A/B testing dashboards all speak in z. This P-value to Z-score Calculator reverses the conversion and returns the critical z behind your alpha. Statisticians, students, and experiment designers reach for it whenever they need boundary values instead of tail areas.
How the conversion works
The calculator inverts the standard normal CDF. For a one-tailed p it solves z = Phi^-1(1 - p); for two tails it splits the risk first, z = Phi^-1(1 - p/2). Example: a two-tailed p of 0.05 maps to 1.96, while a one-tailed 0.05 maps to about 1.64.
Steps to get your z score
Type your P-value into the box, then set Type of P-value to One-tailed or Two-tailed to match the original test. Press Calculate Z-score and read the boundary. Confirm the tail choice carefully, because it is the most common error here, shifting the answer between 1.96 and 1.64 at the same five percent level.
A concrete A/B testing walk-through shows the flow. Your dashboard reports p = 0.02 for a checkout button test and you need the z boundary for a follow up power calculation. Enter 0.02 as the P-value, set Type of P-value to Two-tailed, and press Calculate Z-score: the answer lands at about 2.33, because half of 0.02 leaves 0.01 in each tail. Had the test been declared one-tailed in advance, the same 0.02 would return roughly 2.05, a reminder that the tail setting, not the p-value, decides the number you get.
Frequently Asked Questions
Why does 0.05 give z of 1.96?
A two-tailed 0.05 leaves 2.5 percent in each tail, and the standard normal curve reaches its 97.5th percentile at exactly 1.96. That makes it the classic critical boundary.
What changes with a one-tailed p-value?
All of the alpha moves into one tail, so the required z drops. A one-tailed 0.05 maps to about 1.645 instead of 1.96, an easier bar to clear.
Can the result be negative?
Yes. One-tailed p-values above 0.5 invert below zero, so a p of 0.84 yields roughly z = -1, marking a cutoff point below the distribution mean.
How large can z scores grow?
Extreme p-values push z constantly higher: 0.0001 one-tailed is about 3.72, and 0.0000001 passes 5.1. There is no cap, but every further step demands far more evidence.
Where is this conversion used?
A/B testing teams convert minimum p-values into detection thresholds, auditors turn reported p-values into effect boundaries, and power analyses use the same inversion during sample size planning.