My Daughter's Practice PSAT Score Said '89th Percentile.' The Number Above It Was a Z-Score of 1.22 — Here's What That Actually Means
My daughter's school ran a practice PSAT the first week back this fall, and the score report that came home had two numbers stacked on top of each other: "89th percentile" in bold, and underneath it, in smaller gray text, "z = 1.22." She understood the first number immediately — top 11% of test-takers, she was thrilled. The second number, she skipped right past. So did I, at first, until I actually sat down and worked out where "89th percentile" had come from. It wasn't a separate measurement. The percentile was the z-score, just translated into a number that reads more naturally. Once I saw the formula behind it, half the "mystery stats" on every score report — SAT, ACT, standardized state testing, even the curve on a college midterm — turned out to be the exact same three-input calculation wearing different labels.
What Is a Z-Score and Why Does a Score Report Show One?
A z-score (also called a standard score) measures how many standard deviations a single value sits above or below the mean of its dataset. A z-score of 0 means the value is exactly average; positive means above average, negative means below; and the size of the number tells you how far — a z-score of 2.0 is twice as far from average as a z-score of 1.0.
Score reports use it instead of just showing your raw number because a raw score alone is meaningless without context. Telling a student they scored 1050 on a practice exam says nothing until you know what everyone else scored. A z-score bakes that context directly into a single number, which is also exactly why testing organizations — the College Board, ACT Inc., and most state education departments — convert raw scores to z-scores (and then to percentiles) before reporting results.
Key Takeaway
A z-score answers one question: "how unusual is this number, given everyone else's numbers?" It's the same statistical tool whether the dataset is exam scores, adult heights, or factory part measurements — only the mean and standard deviation change.
How Do You Actually Calculate a Z-Score?
The formula needs exactly three inputs: your raw score, the mean (average) of the whole group, and the standard deviation of the group.
Here, x is the raw score you're checking, μ (mu) is the population mean, and σ (sigma) is the population standard deviation — a measure of how spread out the scores are. A small σ means everyone scored close together; a large σ means scores were all over the map. Here's the report my daughter brought home:
- Her raw score (x): 1050
- The group's mean score (μ): 920
- The group's standard deviation (σ): 107
Subtract the mean from her score first (130 points above average), then divide by how spread out the whole group's scores were (107 points per "typical" gap). The result, 1.22, means her score sat about 1.22 standard deviations above the mean — which is exactly the number printed under her percentile. You can run the same three inputs through the Z-Score Calculator instantly instead of doing the subtraction and division by hand, and it will also return the associated probability.
💡 Pro Tip
If you don't already know the group's standard deviation, you can't skip straight to a z-score — you have to calculate it first from the full dataset. The Standard Deviation Calculator takes a list of raw values and returns σ directly, and the Mean, Median & Mode Calculator finds μ from the same list.
From Z-Score to Percentile: What "89th Percentile" Really Means
A percentile tells you what share of the comparison group scored at or below a given value. "89th percentile" means roughly 89% of test-takers scored the same as or lower than that score — not that the student answered 89% of questions correctly, which is the single most common mix-up parents make reading these reports.
Under a normal distribution — the symmetric bell-shaped curve that most large standardized test populations approximate — every z-score maps to one specific percentile, because the shape of the curve is fixed. That mapping is what a z-table (or a calculator that runs the same lookup) is doing behind the scenes.
| Z-Score | Approx. Percentile | Meaning |
|---|---|---|
| −2.0 | 2nd | Well below average |
| −1.0 | 16th | Somewhat below average |
| 0.0 | 50th | Exactly average |
| +1.0 | 84th | Somewhat above average |
| +1.22 | 89th | Her actual result |
| +2.0 | 98th | Well above average |
This is the same reason the "68-95-99.7 rule" gets taught alongside z-scores: in a normal distribution, about 68% of values fall within 1 standard deviation of the mean, about 95% fall within 2, and about 99.7% fall within 3. That rule is why a z-score of 1.22 lands so close to the 84th-to-98th range rather than somewhere unpredictable — the middle of the curve is dense with values, and the tails are sparse, so each additional standard deviation buys progressively more percentile ground.
Is a Higher Raw Score Always a Better Z-Score?
No — and this is where the "which score is actually better" arguments start, whether it's two kids comparing report cards or two applicants comparing test administrations. A raw score only tells you the position relative to that specific group and that specific spread.
Say a classmate took a different practice exam that same week — a slightly easier version, where more students scored well and the group was more tightly clustered:
| Student | Raw Score | Group Mean | Group SD | Z-Score | Percentile |
|---|---|---|---|---|---|
| My daughter | 1050 | 920 | 107 | 1.22 | ~89th |
| Classmate | 1080 | 1000 | 130 | 0.62 | ~73rd |
The classmate's raw score is 30 points higher — 1080 versus 1050 — but their z-score is lower, because their exam's average was higher and its spread was wider. Relative to their own group, 1080 wasn't nearly as unusual a result as 1050 was relative to mine. A z-score is only ever meaningful next to the specific mean and standard deviation it was calculated from — comparing two raw scores across two different tests or two different years, without converting both to z-scores first, is comparing numbers that don't actually share a scale.
⚠️ Note
Z-scores and percentiles assume a roughly normal (bell-curve) distribution. Small classes, unusual test populations, or heavily skewed data (like income or home prices) don't always follow that shape — the percentile lookup gets less reliable the further the real data strays from a normal curve.
This is also exactly how curved grading works in college. If a professor sets the curve so a B corresponds to a z-score between 0 and +1, a student with the exact same raw score can land a B one semester and a B− the next, purely because that semester's class average and spread shifted. The Grade Calculator handles the separate, more common question of what score you need on a specific assignment to hit a target grade — useful once you already know how the grading scale itself is set.
What's the Difference Between a Z-Score and Standard Deviation?
They're related but answer different questions, and mixing them up is the second most common confusion after "percentile means percent correct."
Standard deviation (σ) measures how spread out an entire dataset is — one number that describes the whole group. A z-score measures where one specific value sits within that spread — a different number for every individual data point, calculated using that group's standard deviation as the yardstick. Standard deviation is the ruler; a z-score is a single measurement taken with that ruler.
You calculate standard deviation once, from the full set of scores. You then calculate a z-score for each individual score you want to evaluate, reusing that same standard deviation every time. If a new student's score comes in later, you don't need to recompute σ — you just plug their raw score into the same formula against the mean and SD you already have.
Key Takeaway
Standard deviation describes the group. A z-score describes one person's (or one measurement's) place inside that group. You need the first to calculate the second.
The same three-input structure shows up well beyond test scores, too. Researchers use z-scores to flag outliers in lab data, quality-control engineers use them to catch a part that's manufactured outside spec, and the Confidence Interval Calculator leans on the same normal-distribution math to estimate a likely range for an unknown population value from a sample.
Frequently Asked Questions
Can a z-score be negative?
Yes — a negative z-score simply means the value is below the mean. A z-score of −1.5 is 1.5 standard deviations below average, which corresponds to roughly the 7th percentile. Only the sign changes; the formula and interpretation work exactly the same way as a positive z-score.
Does a 90th percentile mean you got 90% of questions right?
No — this is the single most common misreading of a score report. A percentile compares you to other test-takers, not to a total possible score. It's entirely possible to be in the 90th percentile with a raw score well below 90% correct, if the exam was difficult enough that most people scored low.
What counts as a "good" z-score?
There's no universal cutoff — it depends entirely on context. A z-score of +1.0 (about the 84th percentile) is strong on a standardized test but might be unremarkable if you're only comparing against a small, high-achieving class. Statisticians commonly flag |z| > 2 or |z| > 3 as an outlier worth a second look, but "good" for a score report is really whatever percentile range the institution reading it considers competitive.
Do you use the population or sample standard deviation for a z-score?
Use the population standard deviation when you have data for the entire group you care about (like every test-taker in a national testing cohort). If you only have a sample — say, one class as a stand-in for all students nationally — statisticians typically use the sample standard deviation (dividing by n−1 instead of n) to correct for the fact that a sample tends to slightly underestimate the true spread.
Why do z-scores and percentiles both appear on the same report?
Because they serve different readers. The percentile is easier for a student or parent to interpret at a glance ("89th percentile" needs no explanation), while the z-score is the more useful number for anyone doing further statistical analysis — comparing across test administrations, tracking a cohort's trend over multiple years, or feeding the data into another formula.
Try It Yourself
A z-score isn't a mysterious extra statistic buried under your percentile — it's the exact calculation that produced the percentile in the first place, just before the final lookup step. Once you know the raw score, the mean, and the standard deviation, you're one subtraction and one division away from understanding exactly where that number stands.
Use the Z-Score Calculator to convert any raw score into a z-score and probability instantly.
Also worth checking:
- Standard Deviation Calculator — find σ from a full list of raw scores
- Mean, Median & Mode Calculator — find the average and center of any dataset
- Confidence Interval Calculator — estimate a likely range for a population value from a sample
- Grade Calculator — figure out what score you need on an assignment to hit a target grade
