How to Interpret Z-Scores: What the Number Actually Means
Computing a z-score takes ten seconds. Knowing what to do with it is the part students actually get marked on — and it's where most explanations wave their hands. So let's be concrete.
Quick answer: the sign of a z-score tells you whether a value is above (+) or below (−) the mean. The size tells you how far out it is. Between −1 and +1 is everyday territory, ±2 is unusual, and ±3 is rare enough to double-check.
Read the sign first
A positive z-score means the value sits above the mean. Negative means below. Zero means dead-on average.
That's it, and yet it's the step people skip. A student once showed me a z of −1.8 for reaction time and panicked because "negative sounds failed." For reaction time, below the mean means faster than average — he had one of the best results in the lab group. Before you judge any z-score, ask two questions: which direction is good for this measurement, and how big is the number? Our post on the negative z-score digs into this properly.
Then read the size: the bands that matter
Here's the interpretation table I wish textbooks printed bigger:
| Z-score | Percentile | How to read it |
|---|---|---|
| 0 | 50th | Exactly average |
| +0.5 | 69.15th | A bit above average |
| +1.00 | 84.13th | Clearly above average |
| +1.50 | 93.32nd | Well above average |
| +1.96 | 97.50th | Unusual — the classic significance cutoff |
| +2.00 | 97.72nd | Unusual; about 2 in 100 score higher |
| +3.00 | 99.87th | Rare — check the data before celebrating |
| −1.00 | 15.87th | Clearly below average |
| −2.00 | 2.28th | Unusual on the low side |
Mirror images hold on the negative side because the bell curve is symmetric. If you need a percentile for a z that isn't in the table, the z score to percentile calculator gives it instantly.
Notice something in that table: the percentiles bunch up near the top. Going from z = +2 to z = +3 only moves you from the 97.72nd to the 99.87th percentile, but the share of people above you shrinks from 2.28% to 0.13% — roughly 17 times rarer. Small z differences in the tails are a much bigger deal than the same difference near the middle.
So what counts as a "good" z-score?
It depends entirely on context, and anyone who gives you a single answer is guessing. Three scenarios:
An exam. z = +1 (84th percentile) is a result most students would take. z = +2 is exceptional. But if the whole class found the test easy and scores bunched at the top, the standard deviation shrinks — and a modest raw gap can inflate into a big z. That's a real quirk of small, bunched-up classes.
A factory measuring bottle fill weights. Here "good" means z near 0. A z of +2.8 on fill weight means that bottle is far off target and worth pulling from the line. Quality teams often flag anything beyond ±3.
Research. "Good" isn't the word at all. Researchers ask whether an effect's z-score clears a threshold like ±1.96, which marks the outer 5% of the distribution (2.5% in each tail). Our critical z score calculator lists those cutoffs for common confidence levels.
The quick gut-check: the empirical rule
When you don't have a table handy, the empirical rule gives you a fast read: about 68% of values fall within ±1, about 95% within ±2, and about 99.7% within ±3. So if someone reports a z of 2.5, you already know — without computing anything — that fewer than 5% of values sit that far out on that side. (The exact percentile for +2.5 is the 99.38th, if you want precision: run it through the z score calculator.)
One last habit worth building: always say the direction out loud when you report a z-score. "1.4 standard deviations above the class mean" is a complete interpretation. "The z-score is 1.4" is just a number waiting to be misread.
Frequently asked questions
What z-score range is considered normal?
For roughly bell-shaped data, z-scores between −1 and +1 cover the middle 68% of values and are as ordinary as it gets. Between −2 and +2 covers about 95%, so anything in there is still unremarkable. Statisticians start paying attention beyond ±2 and treat values past ±3 as rare.
Is a z-score of 0 good or bad?
Neither. A z-score of 0 means the value is exactly average — the 50th percentile. Half the group did better and half did worse. Whether "exactly average" is satisfying depends on what you're measuring and what you were aiming for, not on the number itself.
What z-score is statistically significant?
In a two-tailed test at the common 5% significance level, a z-score beyond ±1.96 is statistically significant. For a stricter 1% level, the cutoffs are ±2.576. These thresholds come from tail probabilities: beyond ±1.96 lies exactly 5% of the distribution, split between the two ends.
Can I compare z-scores from different tests?
Yes — that's one of their main uses. A z of +1.5 in math and +0.5 in English tells you the math performance was relatively stronger, even if the raw scores and maximum marks were totally different. The comparison is fairest when both sets of scores are roughly bell-shaped.