🔍

Understanding P-Values in Hypothesis Testing

Aug 5, 2025

Overview

This lecture introduces the concept of p-value in hypothesis testing, explaining its definition, interpretation, and its relationship to sampling variability and the null hypothesis.

Review of Hypothesis Testing

  • Hypothesis tests involve a null hypothesis (statement with equality) and an alternative hypothesis (statement without equality).
  • The test statistic measures how far sample data deviates from the null hypothesis.
  • Sampling variability means that random samples usually differ from the population and the null hypothesis.

Introduction to P-Value

  • P-value is a tool to help address disagreements between sample data and the null hypothesis.
  • The core question is: Why does my sample data disagree with the null hypothesis?
  • Two possible reasons: the null hypothesis is wrong, or sampling variability caused the disagreement.

Definition and Meaning of P-Value

  • P-value is the probability of obtaining the sample statistic or more extreme results, by sampling variability, assuming the null hypothesis is true.
  • It is a conditional probability calculated under the assumption that the null hypothesis is true.
  • "More extreme" means data further from the null hypothesis than the observed sample statistic.
  • P-values are calculated using the sample statistic, such as sample mean, sample proportion, or test statistic.

Interpreting P-Values

  • A low p-value (close to zero) suggests it is unlikely the data are due to sampling variability, so we reject the null hypothesis.
  • Rejecting the null hypothesis implies that the null is probably wrong, although it is not certain.
  • A higher p-value (not close to zero, e.g., 20%) means the data could plausibly be due to sampling variability.
  • P-value interpretation is always about the null hypothesis, not the alternative.

Key Terms & Definitions

  • Null Hypothesis (H₀) — a statement about the population with an equality claim.
  • Alternative Hypothesis (H₁) — a statement about the population that disagrees with the null.
  • Test Statistic — a value that measures the deviation of the sample from the null hypothesis.
  • Sampling Variability — the natural difference between random samples and the population.
  • P-value — the probability of getting the observed sample statistic or more extreme results if the null hypothesis is true.

Action Items / Next Steps

  • Review the definition and logic of p-value.
  • Prepare for the next lesson on how to calculate p-values.