A z score table is a printed grid of pre-calculated probabilities for the standard normal distribution — the bell curve with a mean of 0 and a standard deviation of 1. Search for the same grid and you will land on it under half a dozen names: a z-score table, a standard normal z-table, a normal distribution table, a z normal distribution table, a table of z-score values, or occasionally “z table normal” typed straight into a search box. They are all the same table of normal distribution probabilities, just labeled differently by whichever textbook printed them. The one skill that actually matters is reading the correct cell: matching the right row to the right column, knowing which tail the number describes, and not losing the sign on a negative z. This guide walks through that mechanics in detail, flags where even careful readers get burned, and ends with a calculator that gives you the exact value without rounding to the nearest table entry.
Key takeaways
| Point | Details |
|---|---|
| Every cell is a left-tail probability | A z score table reports P(Z ≤ z) by default — the share of the standard normal distribution at or below your z-value, not above it. |
| Row holds tenths, column holds hundredths | Row 1.9 plus column 0.06 gives z = 1.96; adding the row and column labels back together should always reproduce the z you looked up. |
| Two incompatible table formats exist | A cumulative table runs from 0 to 1; a 0-to-z table runs from 0 to about 0.5. For a positive z, mixing them up produces an answer exactly 0.5 off. |
| Negative z uses the same mechanics | Look up the negative row directly, or use symmetry — P(Z ≤ −z) = 1 − P(Z ≤ z) — when your table only lists positive values. |
| The table assumes σ is known | Once σ is estimated from a small sample rather than known outright, the t-distribution — not the z score table — is the statistically correct model. |
Quick Checklist: Reading a Z Score Table Without Errors
Run through this before trusting any lookup, whether it comes from a printed normal distribution table or a calculator you have not used before.
Quick checklist: reading a z score table
- Standardize first Convert your raw value to z = (x − μ) / σ before you open any table of normal distribution values.
- Split z into row and column The whole number plus the first decimal is the row; the second decimal is the column.
- Check the sum Row label plus column label should equal the z you are looking for — if it does not, you are in the wrong cell.
- Confirm which tail you have Most modern tables give the left tail, P(Z ≤ z); subtract from 1 for the right tail.
- Identify the format A cell near 0.5 for a small positive z usually signals a 0-to-z table, not a cumulative one.
- Watch the sign A negative z needs the negative row, or the symmetry formula if your table only prints positive values.
Treat this as a pre-flight check, not a memory test — the next sections cover why each step matters.
Calculate a Z-Score Probability Instantly
Enter any z-score below to get the exact cumulative probability, right-tail probability, and area between 0 and z, with full decimal precision instead of the two-decimal rounding a printed table forces on you.
For a larger version with the same options, open the standalone z-table calculator, or browse every probability tool on the calculators hub.
What Is a Z Score Table?
A z score table is a lookup grid built for one specific distribution: the standard normal distribution, written N(0, 1), with mean 0 and standard deviation 1. Every normal distribution table — cumulative or otherwise — describes that one curve, because NIST’s e-Handbook of Statistical Methods defines the standard normal as the case where the general normal density’s location parameter μ equals 0 and its scale parameter σ equals 1.
The reason a printed grid exists at all is that the standard normal’s cumulative distribution function has no closed-form algebraic solution — it has to be computed numerically, integral by integral, point by point. Before cheap computing, someone had to do that numerical work once and publish the results, which is exactly what every distribution z table and normal dist z table in a stats textbook is: a pre-computed answer key so students and analysts did not have to integrate by hand.
A table cell always holds a probability tied to a specific z: in the standard cumulative format, that is P(Z ≤ z), read “the probability that a standard normal variable Z is at or below z.” For z = 1.96, that probability is 0.9750 — 97.5% of the curve sits at or below that point. The underlying bell-curve shape, the 68-95-99.7 empirical rule, and the mean/standard-deviation formula that produces a z-score in the first place are covered in full in the normal distribution guide; this page picks up right where that one hands off — at the table itself.
How Do You Read a Z Score Table?
You read a z score table by splitting your z-value into two parts and finding where they cross. The row carries the whole number plus the first decimal place; the column carries the second decimal place; the cell where they meet is the cumulative probability.
| Desired z | Row | Column | Table value P(Z ≤ z) |
|---|---|---|---|
| 1.96 | 1.9 | 0.06 | 0.9750 |
| −0.83 | −0.8 | 0.03 | 0.2033 |
| 2.14 | 2.1 | 0.04 | 0.9838 |
| −1.25 | −1.2 | 0.05 | 0.1056 |
Walk through one in full. Suppose adult heights in a population are normally distributed with mean μ = 70 inches and standard deviation σ = 3 inches, and you want the share of people shorter than 75.88 inches.
z = (75.88 − 70) / 3 = 5.88 / 3 = 1.96
- Locate the row labeled 1.9.
- Move across to the column labeled 0.06.
- Read the cell: 0.9750.
That means 97.5% of this population is shorter than 75.88 inches — and only 2.5% is taller. The same lookup works for a table of z-score values printed in any format, as long as you know whether the table gives you a left-tail or a 0-to-z number (the next two sections cover both).
How Do You Find Right-Tail and Between-Two-Z Probabilities?
Most z score tables default to the left tail, so anything else requires one extra step of arithmetic. For the area above a value, subtract the left-tail value from 1, because the total area under the standard normal curve equals 1:
P(Z > z) = 1 − P(Z ≤ z)
For z = 1.96, that is 1 − 0.9750 = 0.0250 — the 2.5% right-tail region that forms each tail of a two-tailed test at α = 0.05, since 0.05 split evenly leaves 0.025 on either side.
For the area between two z-scores, subtract the smaller cumulative probability from the larger:
P(z₁ ≤ Z ≤ z₂) = P(Z ≤ z₂) − P(Z ≤ z₁)
Between z = 0.50 and z = 1.50: P(Z ≤ 1.50) − P(Z ≤ 0.50) = 0.9332 − 0.6915 = 0.2417. About 24.17% of the standard normal distribution sits between those two points — two lookups and a subtraction is the entire method.
A special case of that same formula handles symmetric two-tail probabilities, using the curve’s symmetry about 0:
P(−z ≤ Z ≤ z) = 2 × P(Z ≤ z) − 1
For z = 1.96: 2 × 0.9750 − 1 = 0.9500. Exactly 95% of the standard normal distribution falls within ±1.96 — the fact behind the standard 95% two-sided confidence interval, which NIST describes as bracketing the true population parameter in about 95% of repeated samples at that confidence level. The confidence interval guide covers that construction in depth.
How Do You Look Up a Negative Z-Score?
You read a negative z-score exactly the way you read a positive one — find the negative row, then the matching column — but it is where most lookup errors actually happen, because it is easy to drop the sign without noticing.
Take z = −1.25: row −1.2, column 0.05, cell 0.1056. About 10.6% of the standard normal distribution lies at or below −1.25, putting that value near the 11th percentile.
If your particular normal distribution table only lists positive z-values, symmetry gives you the negative-side answer without a second table:
P(Z ≤ −z) = 1 − P(Z ≤ z)
This holds because the curve is a perfect mirror image around 0: the area below −z equals the area above +z, which is one minus the area below +z. It is also why a symmetric two-tailed test only ever requires looking up the positive critical value — the negative one is already implied.
What’s the Difference Between a Cumulative Z-Table and a 0-to-Z Table?
Not every z score table reports the same number for the same z, and the two formats in circulation are easy to confuse because they share the same row-and-column layout.
| Format | What each cell holds | Range | Where it shows up |
|---|---|---|---|
| Cumulative (left-tail) | P(Z ≤ z) | 0 to 1 | Most modern statistics textbooks; the calculator above |
| 0-to-z (body) | P(0 ≤ Z ≤ z) | 0 to about 0.5 | Engineering references and some classic tables, including NIST’s own |
That second row is not a rare exception — NIST’s own cumulative distribution function table for the standard normal distribution is published in the 0-to-z format. NIST’s worked example looks up z = 1.53 by reading row 1.5, column 0.03, for a table value of 0.43699 — the area between 0 and 1.53, not the full area below it. Converting that 0-to-z value to the cumulative probability used throughout this page just means adding 0.5, since everything below 0 already accounts for half the curve:
Left-tail value = 0.5000 + body value
0.5000 + 0.43699 = 0.93699
NIST’s own page uses the same addition for a between-two-points example: P(−1 ≤ Z ≤ 0.5) = (0.5 + 0.19146) − (1 − (0.5 + 0.34134)) = 0.69146 − 0.15866 = 0.5328, matching the between-z method from the previous section. Whichever normal distribution tables you are handed, check the heading or footnote before trusting a cell — for a positive z, a 0-to-z reading used as if it were cumulative lands exactly 0.5 away from the right answer (for a negative z, the cumulative value is 0.5 minus the body value instead).
When Should You Use a T-Table Instead of a Z-Table?
Use a t-table, not a z score table, whenever the standard deviation in your z-formula is estimated from your own sample rather than known from an outside, trusted source. The z = (x − μ) / σ formula assumes σ is a true population parameter; plugging in a sample standard deviation s and reading a z-table understates how uncertain that estimate really is.
NIST’s guidance on confidence limits for the mean builds its interval directly around this distinction, using the t-distribution with N − 1 degrees of freedom whenever the sample standard deviation s stands in for an unknown σ. The common rule of thumb is to lean on the t-distribution whenever n is small (textbooks commonly cite n < 30 as the threshold worth worrying about), because the t-distribution has heavier tails that widen the interval to compensate for the extra uncertainty in s.
That gap closes as the sample grows. NIST’s table of critical values for the Student’s t-distribution lists a row labeled “Normal Values” showing 1.282, 1.645, 1.960, 2.326, 2.576, and 3.090 — the exact z-table critical values this page already uses — because as degrees of freedom climb toward infinity, the t-distribution’s critical values converge on the standard normal’s. By around 100 degrees of freedom the two are already nearly identical, which is why some software defaults to the t-distribution even for large samples: it is never meaningfully wrong, only slightly more conservative. For the full walkthrough of when each test applies, see the t-test vs. z-test comparison and the dedicated t-test guide.
How Do You Get These Probabilities in Excel, Python, or R Instead of a Table?
Every major tool replaces the printed grid with a single function call, and all three return the cumulative left-tail probability by default — the same number this page’s table does.
Excel: =NORM.S.DIST(1.96, TRUE) -> 0.97500
Python: from scipy.stats import norm
norm.cdf(1.96) -> 0.97500
R: pnorm(1.96) -> 0.97500
Excel’s NORM.S.DIST function takes the z-value and a logical cumulative argument; set it to TRUE for the cumulative distribution function, the direct replacement for a table lookup. Python’s SciPy scipy.stats.norm object exposes the same cumulative distribution function through its .cdf() method, with no table or interpolation involved. R’s pnorm() function defaults to mean = 0 and sd = 1 — the standard normal — and accepts a lower.tail = FALSE argument that returns the right-tail probability directly, skipping the 1 − p subtraction a printed table always requires.
A Worked Example: Checking a Machined Part Against a Spec Limit
A machine shop cuts steel shafts meant to measure 25.00 mm in diameter. The process is stable and normally distributed, with mean μ = 25.00 mm and standard deviation σ = 0.08 mm, typical repeatability for this kind of CNC work. An inspector measures one shaft at 24.90 mm. Is that a sign the process has drifted?
- Standardize the reading. z = (24.90 − 25.00) / 0.08 = −0.10 / 0.08 = −1.25.
- Look up z = −1.25. Row −1.2, column 0.05 → 0.1056.
- Interpret it. About 10.56% of shafts from this process run this small or smaller — roughly the 11th percentile. On its own, that is an unremarkable reading from a process behaving normally, not evidence of drift.
Now suppose the part has a hard lower spec limit of 24.84 mm — anything smaller gets scrapped. What fraction of total output falls below that line?
z = (24.84 − 25.00) / 0.08 = −0.16 / 0.08 = −2.00
Row −2.0, column 0.00 → 0.0228.
- Decide. About 2.28% of shafts from this stable process will land outside the lower spec limit by chance alone, with nothing wrong on the machine — a scrap rate worth flagging to engineering before it becomes a customer complaint. That is the real payoff of the lookup: it turns one measurement into a statement about the whole production run.
Common Mistakes When Reading a Z Score Table
- Mixing up rows and columns. Reading the column as the tenths digit instead of the hundredths digit turns row 1.9, column 0.06 (z = 1.96) into row 1.6, column 0.09 (z = 1.69) — a completely different value. Fix it by checking that row plus column always equals your target z.
- Forgetting the table gives left-tail values. Asked for the probability above a point, skipping the 1 − P(Z ≤ z) step turns a correct 0.0250 right-tail answer into an incorrect 0.9750. Fix it by always asking “which side of the line do I actually need?” before reading off a cell.
- Treating a 0-to-z table as cumulative. A body-format cell of 0.43699 is not the probability of falling below that z — it is 0.5 short. Fix it by checking the table’s heading or footnote for its format before trusting any number from it.
- Losing the sign on a negative z. Looking up z = 1.25 when the real value is z = −1.25 swaps the 11th percentile for the 89th. Fix it by writing the sign down before you open the table, not after.
- Rounding the z-score too early. Rounding to one decimal place instead of two moves the lookup further than it looks — P(Z ≤ 2.00) = 0.9772 versus P(Z ≤ 1.96) = 0.9750, a difference of about 0.002 from dropping just the hundredths digit. Fix it by keeping two decimal places through the lookup, or using the calculator above for exact precision.
- Applying the table to non-normal data. Standardizing a strongly skewed variable — hold times, income, reaction times — and reading the result off a z score table produces a confident-looking number that does not describe the real distribution. Fix it by checking a histogram or a skewed-distribution diagnosis before you standardize anything.
Statohub’s Take on the Z Score Table
Statohub’s position is that the z score table is worth learning by hand once, even though almost nobody looks one up by hand anymore. Working a few lookups the slow way builds an intuition for how probability concentrates around the mean that a calculator’s instant answer does not teach on its own. Once that intuition is in place, reach for the calculator above or a software function for anything that actually matters — the two-decimal rounding a printed grid forces is fine for a classroom exercise and not fine for a real decision.
Put the Z Score Table to Work With Statohub’s Tools
The calculator above handles the full lookup — left-tail, right-tail, and between-z probabilities — without any row-and-column hunting. For a standalone version, open the z-table calculator directly. The Probability & Distributions hub covers the surrounding distributions this table connects to, including the discrete binomial distribution and the exponential distribution, and the probability formula guide covers the general rules every one of these probabilities still has to obey. Browse every statistics tool on Statohub’s calculators hub when you are ready to check your own numbers.
Sources
Sources
- NIST/SEMATECH e-Handbook of Statistical Methods — 1.3.6.6.1. Normal Distribution NIST
- NIST/SEMATECH e-Handbook — 1.3.6.7.1. Cumulative Distribution Function of the Standard Normal Distribution NIST
- NIST/SEMATECH e-Handbook — 1.3.6.7.2. Critical Values of the Student’s-t Distribution NIST
- NIST/SEMATECH e-Handbook — 1.3.5.2. Confidence Limits for the Mean NIST
- NIST/SEMATECH e-Handbook — 7.1.4. What Are Confidence Intervals? NIST
- scipy.stats.norm — SciPy v1.18.0 Manual SciPy
- NORM.S.DIST function Microsoft Support
- R: The Normal Distribution (dnorm, pnorm, qnorm, rnorm) The R Project
FAQ
Frequently asked questions
- What Is a Z Score Table Used For?
- A z score table converts a standardized z-score into a cumulative probability — the fraction of the standard normal distribution at or below that value. That covers finding p-values for z-tests, locating critical values for confidence intervals, computing percentile ranks in a normally distributed population, and solving normal-probability problems without integrating by hand.
- How Do I Find the Probability for a Z-Score Not Listed in a Normal Distribution Table?
- Round to two decimal places and use that entry — for z = 1.963, look up 1.96. For that specific rounding, the error is only about 0.00017 in cumulative probability (P(Z ≤ 1.963) = 0.97518 versus the table’s 0.97500), which is fine for coursework. For exact precision, use the calculator above, which computes the probability directly for any z-score rather than rounding to the nearest grid cell.
- What Does a Table Value of 0.9750 Mean?
- It means 97.5% of the standard normal distribution falls at or below that z-value — equivalently, a random draw from N(0, 1) is less than or equal to z with probability 0.9750, leaving 2.5% in the right tail. That figure corresponds to z = 1.96, the critical value behind a two-sided 95% confidence interval.
- What Z-Score Corresponds to the 95th Percentile?
- Search the table of z-score values closest to 0.9500. P(Z ≤ 1.64) = 0.9495 and P(Z ≤ 1.65) = 0.9505, so interpolating gives the 95th-percentile z-score as approximately 1.645, usually rounded to 1.64 or 1.65. That one-tailed figure is different from the two-tailed 95% critical value of ±1.96, which marks off the middle 95% of the distribution rather than its bottom 95%.
- How Do I Use a Z Score Table for a Hypothesis Test?
- For a one-tailed z-test at significance level α, find the z where P(Z ≤ z) equals 1 − α. At α = 0.05, that target is 0.9500, giving z ≈ 1.645; reject the null hypothesis if your test statistic exceeds it. For a two-tailed test at α = 0.05, find z where P(Z ≤ z) = 0.9750, which is z = 1.96, and reject if the absolute value of your test statistic exceeds that.
- Why Do Some Normal Distribution Tables Give Different Numbers for the Same Z?
- Because two incompatible formats circulate: a cumulative table reports P(Z ≤ z) directly, while a 0-to-z (body) table reports only the area between 0 and z, which runs at most to about 0.5. Converting between them just means adding or subtracting 0.5 for a positive z — but skipping that conversion is why two people can read the same cell and land 0.5 apart on the answer.