What a Z Score Means and How to Calculate It Correctly

What Is a Z Score?

A z score is a way of describing exactly where a single value sits within a data set, using the mean and standard deviation as reference points. Instead of just knowing a raw number, like a test score of 87 or a height of 68 inches, a z score tells you how that number compares to everyone else in the group it came from.

The idea is simple once you see it in context. Standard deviation measures how spread out a data set is overall. A z score takes that same measurement and applies it to one specific value, expressing the distance between that value and the mean in standard deviation units, rather than in whatever units the data happens to be measured in.

That last part is what makes z scores so useful. Once a value is converted into a z score, the original units stop mattering. A z score of 1.5 means the same thing whether it came from a test score, a height measurement, or a company's quarterly revenue: the value sits one and a half standard deviations above average.

Why Z Scores Matter

Z scores solve a specific, common problem: comparing values that don't share the same scale. Say two employees are reviewed on completely different metrics — one's performance is measured in dollars of sales, the other's in a customer satisfaction percentage. Comparing $85,000 in sales to a 92% satisfaction score directly tells you nothing useful.

Convert both to z scores against their own team's average and standard deviation, though, and you're suddenly comparing apples to apples: whoever has the higher z score is performing further above their team's typical range, regardless of the original units.

That standardizing power is why z scores show up across so many fields, from education to healthcare to finance. Here's what regular use of z scores gives you:

• A standardized way to compare values from completely different scales or units
• A quick gauge of whether a single result is typical or unusual, without eyeballing the whole data set
• The foundation for percentiles, which are often easier to explain than raw z scores
• A building block for hypothesis testing and other statistical methods used throughout research
• A practical way to flag data entry errors or genuine outliers before they skew an analysis

Once you're comfortable with the basics here, our guide on how to use z scores to compare different data sets goes further into that standardization use case specifically.

The Z Score Formula Explained

The z score formula is short, but each piece of it matters:

To use it, subtract the mean from your value, then divide the result by the standard deviation. That's the entire calculation.

If you're working with a full population, μ and σ represent the population mean and population standard deviation. If you're working with a sample instead, which is far more common in practice since you rarely have data on an entire population, the same formula works using the sample mean (often written x̄) and sample standard deviation (s) in place of μ and σ. Our guide on mean absolute deviation vs standard deviation covers how standard deviation itself gets calculated, or 100 Calculator's Mean Absolute Deviation Calculator can help you check your spread numbers first.

The result, z, can be positive, negative, or exactly zero, and it's almost always expressed in standard deviation units rather than the original units of the data.

How to Calculate a Z Score Step by Step

Here's the full process broken into steps, followed by a worked example with real numbers.

1. Find the mean (μ) of your data set. Add up every value, then divide by how many values there are.
2. Find the standard deviation (σ). This measures how spread out the values are around that mean. If you don't already have it, you'll need to calculate it before moving forward.
3. Subtract the mean from your value: x − μ.
4. Divide that result by the standard deviation: (x − μ) / σ.
5. The number you get is the z score for that value.

Worked Example

Say a class takes a math test. The average score comes out to 75, with a standard deviation of 8. One student scores 87. Here's how to find that student's z score:

z = (87 − 75) / 8
z = 12 / 8
z = 1.5

That student's score sits 1.5 standard deviations above the class average — solidly above typical, but nowhere near outlier territory. If a different student scored 59, their z score would be (59 − 75) / 8 = -2, meaning their score fell two full standard deviations below average.

How to Interpret a Z Score

Once you have a z score, its sign and size tell you everything you need to know. Here's a quick reference for how to read one:

What Counts as an Outlier?

There's no single official cutoff for what makes a value a statistical outlier, but a common rule of thumb treats z scores beyond +3 or -3 as potential outliers, since so little of a normal distribution falls out that far. Some fields, especially those dealing with quality control or safety, use a tighter threshold around ±2.5.

Whether a value truly counts as an outlier also depends on your sample size and whether the underlying data is roughly normally distributed in the first place. A genuinely skewed data set can produce large z scores that aren't outliers at all, just a natural feature of that data's shape.

Z Scores and the Normal Distribution

Z scores are most useful, and most commonly discussed, in the context of a normal distribution: the symmetric, bell-shaped curve that shows up constantly in nature and in large data sets, from heights and test scores to measurement errors and manufacturing tolerances.

When data follows a normal distribution, converting every value to a z score transforms that data onto something called the standard normal distribution: a bell curve with a mean of exactly 0 and a standard deviation of exactly 1. That transformation is what makes z-tables and percentile lookups work consistently across completely different original data sets.

The Empirical Rule (68-95-99.7 Rule)

This predictable spread is often summarized as the empirical rule, sometimes called the 68-95-99.7 rule. The (NIST) e-Handbook of Statistical Methods describes this same pattern: in a normal distribution, about 68% of values fall within 1 standard deviation of the mean, about 95% fall within 2 standard deviations, and about 99.7% fall within 3 standard deviations.

That's a handy shortcut. Once you know a value's z score, you immediately know roughly how common or rare it is, without needing to look anything up. A z score of 2, for instance, means that value is higher than roughly 97.7% of the data, since the 95% figure covers ±2 combined and the remaining 5% splits evenly between both tails.

Positive, Negative, and Zero Z Scores

The sign of a z score is just as important as its size, since it tells you which direction a value differs from average.

A positive z score means the value is above the mean. The larger the positive number, the further above average it sits. A z score of +2 is further above average than a z score of +0.5.

A negative z score means the value is below the mean. The same logic applies in reverse: a z score of -2 is further below average than a z score of -0.5. It's easy to accidentally compare the size of two z scores while ignoring their signs, which is one of the most common interpretation mistakes people make (more on that in a moment).

A z score of exactly zero means the value is precisely equal to the mean, neither above nor below. On a normal distribution curve, zero sits right at the peak of the bell, since that's where values cluster most densely.

For example, if two students each score exactly at their own class's average, both get a z score of 0, even if one class averaged 60 points and the other averaged 90. The z score strips away the original scale entirely and just tells you: this value equals the average, full stop.

Z Score vs Standard Deviation vs Percentile

These three terms get used together so often that it's easy to blur them into one concept, but each one measures something different.

Standard deviation is a property of the whole data set. A z score is a property of one individual value within that data set. A percentile is just another way of expressing that same value's position, translated into a "higher than X% of the group" format that many people find more intuitive than standard deviation units.

Converting between a z score and a percentile requires a standard normal table, often called a z-table, or a calculator, since the relationship follows a specific statistical curve rather than a simple straight-line formula. Our guide on how to interpret a p value in hypothesis testing covers a closely related conversion, since p values in a z-test are calculated directly from a value's z score.

Real-World Uses of Z Scores

Z scores aren't just a classroom exercise. They show up constantly in fields that need a consistent way to flag unusual values.

Standardized Testing and Grading

Test scores are one of the most common places people encounter z scores, even if the report calls it something else, like a "scaled score." When a test is graded on a curve, raw scores often get converted to z scores first, then translated into a final grade or percentile rank, so a "good" score means the same thing regardless of how hard that particular test version happened to be.

Growth Charts in Pediatric Care

Pediatric growth charts use z scores to compare a child's height, weight, and other measurements against a large reference population of children the same age and sex. (CDC) A weight-for-age z score of -2 or lower, for example, is a standard flag that a doctor may look into further, since it means that child's weight sits well below most of their peers.

Quality Control and Six Sigma

Manufacturing and quality control teams use z scores to track how far a product's measurements drift from specification. The "sigma" in Six Sigma literally refers to standard deviations: a Six Sigma process is one where defects sit roughly six standard deviations from the target, making them exceptionally rare.

Finance and Investing

In finance, z scores help identify unusual returns or risk levels by comparing one data point, like a stock's daily return, against its own historical average and volatility. An unusually large z score on a given day can signal an event worth investigating, rather than typical day-to-day noise.

Scientific Research and Hypothesis Testing

Researchers use z scores as the backbone of the z-test, a statistical method for deciding whether an observed result is likely due to chance or reflects a real effect. The z score calculated during a z-test gets converted into a p value, which is what most published research actually reports. Our guide on what a p value really tells you about significance breaks down that connection in more detail, and 100 Calculator's P Value Calculator can help once you're ready to take that next step.

Common Mistakes to Avoid

A z score calculation is simple arithmetic, but a few small missteps can quietly throw off the result or the interpretation. Watch for these:

• Mixing up population and sample standard deviation, which produces slightly different results, especially with small data sets
• Ignoring the sign and treating a z score of -2 the same as +2, when they actually point in opposite directions
• Assuming z scores always translate cleanly into percentiles, when that relationship only holds accurately for roughly normal distributions
• Rounding intermediate steps too early, which can shift the final z score more than expected on borderline values
• Confusing a z score with a percentile and using the two terms interchangeably in a report or analysis
• Using a mean and standard deviation from the wrong reference group, which produces a technically correct but meaningless z score

Most of these are easy to catch with a quick sanity check: does the sign make sense given whether the value is above or below the group's typical range? Does the reference mean and standard deviation actually match the group you're comparing against? A few seconds of double-checking catches the majority of z score errors before they make it into a final report.

Using a Z Score Calculator

Calculating a z score by hand is simple enough for a single value, but it gets tedious fast if you're checking several values against the same data set, or if you'd just rather skip the arithmetic entirely.

100 Calculator's Z Score Calculator takes your raw value, mean, and standard deviation, and instantly returns the z score along with the percentile it corresponds to. It handles both population and sample calculations, so you don't need to remember which version of the formula applies to your situation.

It's especially useful for double-checking a manual calculation, working through multiple values quickly, or converting a z score back into a percentile without pulling up a full z-table. Like every tool on 100 Calculator, it's free to use and doesn't require an account.

Sources & References

This guide draws on publicly available statistics references, including:

• National Institutes of Health, National Library of Medicine — an overview of z scores and how they're used in health statistics
• NIST/SEMATECH e-Handbook of Statistical Methods — a public reference maintained by the National Institute of Standards and Technology
• CDC Growth Charts — documentation on how the Centers for Disease Control and Prevention uses z scores in pediatric growth tracking

More From Math & Statistics Guides

Still building your statistics toolkit? These related guides dig deeper into concepts that pair naturally with z scores.

Frequently Asked Questions