Mathematics
Z-Score
Quick fact
A z-score of 2 means your data point is two standard deviations above the mean—and in a normal distribution, about 95% of all values fall within that range.
Why this is interesting
If you scored 85% on a test, are you a star student? Only if you know what everyone else scored—the z-score tells you exactly where you stand in the crowd.
Read the full explanation
Understanding Z-Score
Imagine you and your friends are all different heights. We want to see who is unusually tall. The z-score tells us exactly that: it's a number that shows how many 'steps' (standard deviations) a person's height is from the average height. To calculate it, you take the difference between the value and the mean, then divide by the standard deviation. The result is a standard score—usually between -3 and 3. A positive z-score means above average, negative means below average, and zero means exactly average. This standard score lets you compare or rank data that come from different sets. For example, you can compare a math test score with a reading test score, even if the tests have different scoring scales.
A deeper explanation
The z-score's power comes from the properties of the normal distribution. When data is bell-shaped, the z-score maps every value to a position on a standard normal curve, where the mean is 0 and the standard deviation is 1. This is why it's also called standardization. By converting raw data to z-scores, you can use the empirical rule: about 68% of values fall within ±1 z-score, 95% within ±2, and 99.7% within ±3. This directly makes outlier detection possible—anything beyond ±3 is often considered an anomaly. Moreover, z-scores enable the comparison of data from different normal distributions, because you're removing scale and shifting the center to 0. This is the foundation for many inferential statistics, such as z-tests for sample means, and it's also used in quality control and data normalization for machine learning.