Direct comparison
Z-score vs t-score
A z-score uses the known population standard deviation and the standard normal distribution; a t-score is used when the standard deviation is estimated from the sample, relying on the t-distribution and its degrees of freedom.
The step most authors miss
Doing CRediT right? Don’t stop at the statement.
A CRediT statement credits you inside one paper. The recognition CRediT was built for happens when those roles are tied to you, persistently. Sign in with your ORCID — free — and claim your CRediT contributions on casrai.org, the home of the standard. They become a verified, portable part of your identity, not a line that disappears into one PDF.
Free: claim your contributions, then export a journal-ready CRediT statement, schema.org structured data, JATS XML, CSV or BibTeX — and preview your public profile. A membership publishes that profile publicly and verifies the journals you serve.
Side-by-side comparison
| Dimension | Z-score | T-score |
|---|---|---|
| When to use | Population standard deviation σ is known. | Population standard deviation is unknown and estimated by s. |
| Distribution | The standard normal distribution (mean 0, SD 1). | The Student’s t-distribution. |
| Sample size | Suited to large samples (or known σ). | Designed for small samples; especially important when n is small. |
| Degrees of freedom | None — the standard normal has a single fixed shape. | Shape depends on the degrees of freedom (often n − 1). |
| Tail thickness | Thinner tails (the fixed normal curve). | Heavier tails, reflecting extra uncertainty from estimating σ. |
| Formula form | z = (x − μ) / σ for a value; (x̄ − μ) / (σ / √n) for a mean. | t = (x̄ − μ) / (s / √n) for a sample mean. |
| Critical values | Fixed (for example, 1.96 for a two-tailed 95% test). | Larger than z for small df, shrinking towards z as df grows. |
| As sample grows | Already the limiting reference curve. | The t-distribution converges on the normal as df increases. |
| Typical test | Z-test of a mean or proportion with known variance. | One-sample, paired or two-sample t-test of means. |
The same logic under different knowledge
A z-score and a t-score do the same job — expressing how far a value or a sample mean lies from a reference, in standardised units — but they assume different things. The z-score assumes the population standard deviation is known, an assumption rarely met in practice. When the standard deviation must instead be estimated from the sample, that estimation adds uncertainty, which the t-distribution accounts for through its heavier tails. The smaller the sample, the more this matters, so the t-distribution’s shape depends on the degrees of freedom. As the sample grows, the estimate of the standard deviation improves and the t-distribution converges on the standard normal — which is why, for large samples, the two are nearly identical.
Common questions
FAQ
When should I use a t-score instead of a z-score?+
Use a t-score when the population standard deviation is unknown and must be estimated from the sample — which is the usual situation in research — and particularly when the sample is small. Use a z-score when the population standard deviation is known, or when the sample is large enough that the t and normal distributions are practically indistinguishable.
Why does the t-distribution have heavier tails than the normal?+
The t-distribution accounts for the extra uncertainty introduced by estimating the standard deviation from the sample rather than knowing it. This added uncertainty makes extreme values more likely, which thickens the tails. The fewer the degrees of freedom, the heavier the tails; as the sample size grows, the estimate improves and the t-distribution narrows towards the normal.
Do z-scores and t-scores give different answers for large samples?+
For large samples the difference is negligible. As the degrees of freedom increase, the t-distribution converges on the standard normal, so the critical values and resulting conclusions become almost identical. The distinction matters most for small samples, where the t-distribution’s wider critical values give a more honest, more conservative test.
Going deeper
