z* is the number of standard errors you go out from the estimate in a confidence interval, and the cut-off for a z-test. It comes straight from the standard normal z-table: for a two-sided level C, z* is the z with a left area of 1 − α/2.
Tap a confidence level to shade it on the curve. Blue: the confidence area. Red: the tail area α (two-sided: α/2 in each tail). z* is rounded to 3 decimals; the 4-decimal value is shown under the table. Source: the standard normal distribution, the same values as our z-table.
Question (illustrative numbers): a sample of n = 36 has mean 100, and the population standard deviation is known to be σ = 15. Build a 95% confidence interval for the mean.
Step 1. Find z*: 95% two-sided → α = 0.05, α/2 = 0.025, left area 0.975 → z* = 1.960 (row 95% above).
Step 2. Standard error: σ / √n = 15 / √36 = 15 / 6 = 2.5.
Step 3. Margin of error: z* × SE = 1.960 × 2.5 = 4.9.
Answer: 100 ± 4.9, so the interval is (95.1, 104.9). At 99% you would use 2.576 instead, giving 100 ± 6.44.
Pages for the most-used levels: 90% confidence interval z-score, 95% confidence interval z-score and 99% confidence interval z-score.
z* = 1.960 for a two-sided 95% interval or two-tailed test. Each tail holds α/2 = 0.025, so the left area at z* is 0.975, which the z-table shows exactly at z = 1.96. For a one-sided 95% bound it is 1.645.
Two-sided: 1.645 for 90%, 1.960 for 95% and 2.576 for 99%. One-sided: 1.282 for 90%, 1.645 for 95% and 2.326 for 99%.
Compute α = 1 − C. For a two-sided interval, find the z whose left area is 1 − α/2; for one-sided, the z whose left area is C. Read it from the z-table or use the inverse normal calculator. In Excel: =NORM.S.INV(1 − α/2).
One-sided puts all of α in one tail, so you do not need to go as far out. At 95%, one tail of 0.05 gives 1.645, while two tails of 0.025 give 1.960.
Use z* when the population standard deviation is known or for large-sample proportions. When you estimate σ from a small sample, use t* with n − 1 degrees of freedom from the t-table on ttable.org.
Yes. Tap “Print this table” to print just the chart. Choose one-sided or two-sided first, and use “Hide z*” to make a self-test sheet.