A z-score tells you how many standard deviations a value is from the mean. It sounds technical but it's enormously useful — here's a plain-English explanation with practical examples.
A z-score answers one of the most common questions in data analysis: "Is this value unusual, and if so, by how much?" Once you understand z-scores, you can compare values from completely different datasets on the same scale.
A student scores 82 on a test where the class average is 70 and the standard deviation is 8.
Another student scores 58 on the same test:
In a normal distribution, roughly 68% of values have z-scores between −1 and +1. About 95% fall between −2 and +2. Values with z-scores beyond ±3 are genuinely rare — less than 0.3% of data.
This is where z-scores become extremely powerful. Say a student takes two tests on the same day:
The raw scores suggest physics performance was better (65 vs 78). The z-scores reveal the opposite — the history score was more exceptional relative to peers. Z-scores level the playing field between tests with different means and spreads.
Medicine: Growth charts express children's height and weight as z-scores relative to reference populations. A child at z = −2 for height is at the 2nd percentile — a clinical flag for growth monitoring.
Finance: The Altman Z-score predicts corporate bankruptcy risk using financial ratios combined into a composite z-score. A score below 1.81 historically indicated high bankruptcy risk.
Quality control: A machine producing parts with a measured dimension is evaluated by its z-score relative to the specification tolerance. If the process mean drifts, z-scores show exactly how many standard deviations away from specification you are.
Sports analytics: Player performance metrics are converted to z-scores to compare players who played under different team or conditions contexts.
Using a z-table (or statistical software), any z-score can be converted to a probability. A z-score of +1.96 corresponds to the 97.5th percentile — meaning 97.5% of normally distributed values fall below it. This is why 1.96 appears constantly in statistics: a 95% confidence interval extends ±1.96 standard deviations from the mean.