Calculating P-Values from Test Statistics
What you'll learn: How to transform a test statistic into a p-value using probability distributions.
The Connection Between Test Statistics and P-Values
You've already learned what a test statistic is (a standardized measure of how far your sample result is from the null hypothesis) and what a p-value represents (the probability of seeing results as extreme as yours if the null hypothesis is true). Now we connect them.
The test statistic tells you where you landed on a reference distribution. The p-value tells you how unusual that location is.
How the Calculation Works
Every test statistic follows a known probability distribution under the null hypothesis—Z follows a standard normal, t follows a t-distribution, etc. These distributions come with cumulative distribution functions (CDFs) that tell you the probability of getting a value at or below a certain point.
To find the p-value:
- Calculate your test statistic (you learned this in lesson 316)
- Identify the reference distribution (standard normal for Z, t-distribution for t, etc.)
- Use the CDF to find the area beyond your test statistic
For a two-sided test, you calculate the probability in both tails (values as extreme or more extreme in either direction). For a one-sided test, you only care about one tail direction.
Simple Example
If your Z test statistic is 2.5 in a two-sided test, you find the area beyond 2.5 in both directions. The CDF tells you about 99.38% of values fall below 2.5, so 0.62% are above. Double that for both tails: p-value ≈ 0.0124.
Key Takeaway: P-values are calculated by using cumulative distribution functions to measure how far into the tails of the reference distribution your test statistic falls—transforming a number into a probability of extremeness.