Course contentsShow
← Data Science
Lesson 315 of 2,1459. Hypothesis Testing FoundationsFree lesson

Common Test Statistics: Z, t, Chi-Square, and F

Overview of the most frequently used test statistics and the contexts in which each is appropriate.

Common Test Statistics: Z, t, Chi-Square, and F

You'll learn: Which test statistics are used most often in hypothesis testing and when each one applies.

The Four Workhorses of Hypothesis Testing

Once you've calculated a test statistic (lesson 314), you need to know which test statistic to use. Four families cover most scenarios you'll encounter:

Z-Statistic

Use when you know the population standard deviation or when working with proportions. Think of Z as the "ideal world" test—it assumes you have complete information. Common when sample sizes are large (n ≥ 30) and you're testing means or proportions.

Context: "Is this sample mean different from a known population mean?" or "Is this proportion significantly different from 50%?"

t-Statistic

Use when you don't know the population standard deviation and must estimate it from your sample. The t-test accounts for extra uncertainty from estimation, especially critical with smaller samples. This is the most common choice for comparing means.

Context: "Is the average height in this sample different from 5.5 feet?" or "Do these two groups have different average scores?"

Chi-Square (χ²)

Use for categorical data or when testing variance. Chi-square tests relationships between categories or whether observed frequencies match expected patterns.

Context: "Is there a relationship between gender and voting preference?" or "Is this die fair?"

F-Statistic

Use when comparing variances or testing multiple groups simultaneously (like ANOVA). The F-test compares variation between groups to variation within groups.

Context: "Do three different teaching methods produce different average test scores?" or "Are these two population variances equal?"

Key Takeaway: Choose your test statistic based on what you're measuring (means vs proportions vs categories), what you know (population parameters), and how many groups you're comparing—Z for known parameters, t for unknown, Chi-square for categories, and F for multiple group comparisons.