Why t-Tests? When the Population Variance is Unknown
You'll learn when to replace z-tests with t-tests and why the Student's t-distribution matters.
The Problem with z-Tests in Real Life
In previous lessons, you used z-tests to test hypotheses about population means. But there's a catch: z-tests require you to know the population standard deviation σ. In practice, you almost never know σ—you only have your sample data.
When σ is unknown, you must estimate it using the sample standard deviation s. This introduces extra uncertainty because s itself varies from sample to sample. Simply plugging s into the z-test formula and pretending nothing changed would underestimate your uncertainty and lead to overconfident conclusions.
Enter the t-Distribution
The Student's t-distribution accounts for this added variability. When you compute:
this statistic doesn't follow a standard normal distribution—it follows a t-distribution with n - 1 degrees of freedom.
The t-distribution looks like the normal distribution but has heavier tails, meaning extreme values are more likely. This reflects the extra uncertainty from estimating σ. As your sample size grows, s becomes a better estimate of σ, and the t-distribution converges toward the standard normal distribution.
When to Use t-Tests
- Population variance unknown: Use t-tests
- Population variance known (rare): Use z-tests
- Large samples (n ≥ 30): t and z give similar results, but t is still more appropriate
Key Takeaway: Use t-tests when the population variance is unknown and you estimate it from your sample; the t-distribution's heavier tails properly account for the extra uncertainty introduced by estimating σ with s.