Quick Summary
This guide will equip you to master the interpretation of p-values, a cornerstone of statistical inference. You will learn the precise definition of a p-value as a conditional probability and how to use it to make a formal, evidence-based conclusion in a significance test. By the end of this lesson, you will be able to compare a p-value to a significance level (α), decide whether to reject or fail to reject a null hypothesis, and write a clear, contextual conclusion that would earn full credit on the AP exam.
Key Concepts
The p-value is the central piece of evidence in a significance test. Understanding what it is—and what it is not—is critical for making sound statistical conclusions.
1. The Definition of a p-value
The p-value is the probability of obtaining a sample statistic as extreme or more extreme than the one actually observed, assuming the null hypothesis (H₀) is true.
It's a Conditional Probability: The entire calculation is predicated on the assumption that the null hypothesis is true. We temporarily live in a world where H₀ is correct and then calculate the probability of our sample result occurring in that world.
"As Extreme or More Extreme": This phrase is crucial. It means we're not just calculating the probability of getting our specific sample result, but the probability of getting that result or anything even further away from the null hypothesis value. The direction of "extreme" is determined by the alternative hypothesis (Hₐ).
Right-tailed test (Hₐ: p > p₀): "Extreme" means greater than or equal to our observed statistic.
Left-tailed test (Hₐ: p < p₀): "Extreme" means less than or equal to our observed statistic.
Two-tailed test (Hₐ: p \neq p₀): "Extreme" means as far away from the null value in either direction. We find the probability in one tail and double it.
[Image: A normal distribution curve centered at the null hypothesis value. The observed test statistic is marked on the x-axis, and the p-value is the shaded area in the tail(s) beyond that statistic.]
2. The p-value as Evidence
Think of the p-value as a "surprise index." A small p-value indicates that our observed sample result is very surprising or unlikely to have occurred if the null hypothesis were true.
Small p-value: If the p-value is small, it means our sample result is rare in the world where H₀ is true. This leads us to doubt our initial assumption. A small p-value provides strong evidence against the null hypothesis (H₀) and in favor of the alternative hypothesis (Hₐ).
Large p-value: If the p-value is large, it means our sample result is not surprising. It's a plausible, reasonably likely outcome if the null hypothesis were true. A large p-value provides weak evidence against the null hypothesis (H₀). We conclude that we don't have enough evidence to discard it.
3. The Significance Level (α) and The Decision Rule
To avoid subjective interpretations of "small" or "large," we use a pre-determined threshold called the significance level, denoted by α (alpha).
- Significance Level (α): This is the probability of making a Type I error (rejecting a true null hypothesis) that we are willing to risk. It's the standard by which we judge our p-value. Common values are α = 0.05, α = 0.01, and α = 0.10. If α is not stated, assume α = 0.05.
The decision rule is simple and absolute:
If p-value \le α, our result is statistically significant. We reject the null hypothesis (H₀).
If p-value > α, our result is not statistically significant. We fail to reject the null hypothesis (H₀).
4. The Two Possible Conclusions
Based on the decision, we write a conclusion in the context of the problem. There are only two ways to phrase this conclusion.
When we Reject H₀ (p-value \le α):
- "We have convincing statistical evidence that [state the alternative hypothesis, Hₐ, in words]."
When we Fail to Reject H₀ (p-value > α):
- "We do not have convincing statistical evidence that [state the alternative hypothesis, Hₐ, in words]."
Notice we never say we "accept" H₀ or have proven H₀ is true. A large p-value simply means we lack sufficient evidence to reject it, much like a "not guilty" verdict in a trial doesn't prove innocence.
Key Vocabulary
p-value: The probability of obtaining a test statistic as extreme or more extreme than the observed value, assuming the null hypothesis is true.
Significance Level (α): A pre-determined threshold used to decide if a p-value is small enough to be statistically significant. It represents the maximum allowable probability of a Type I error.
Statistically Significant: A result is deemed statistically significant if its p-value is less than or equal to the chosen significance level (α), leading to the rejection of the null hypothesis.
Null Hypothesis (H₀): The initial claim of no effect, no difference, or no change, which is tested for possible rejection.
Alternative Hypothesis (Hₐ): The claim for which we are trying to find evidence; it is accepted if the null hypothesis is rejected.
Reject H₀: The decision made in a significance test when the p-value is less than or equal to the significance level (α).
Fail to Reject H₀: The decision made in a significance test when the p-value is greater than the significance level (α).
Calculator Tech (TI-84)
While this topic is about interpretation, the p-value itself is often calculated from a test statistic (like z or t). If you are given a test statistic and need to find the p-value:
For a z-test statistic (Proportions):
Go to
2nd-> .Select
2: normalcdf(.Inputs:
Left-tailed test (Hₐ: p < p₀):normalcdf(-1E99, your_z_score, 0, 1)
Right-tailed test (Hₐ: p > p₀):normalcdf(your_z_score, 1E99, 0, 1)
Two-tailed test (Hₐ: p \neq p₀): Calculate one tail (usually the one corresponding to your z-score's sign) and multiply the result by 2.
For a t-test statistic (Means):
Go to
2nd-> .Select
6: tcdf(.Inputs:
Left-tailed test (Hₐ: μ < μ₀):tcdf(-1E99, your_t_score, df)
Right-tailed test (Hₐ: μ > μ₀):tcdf(your_t_score, 1E99, df)
Two-tailed test (Hₐ: μ \neq μ₀): Calculate one tail and multiply the result by 2.
Note: Full significance test functions like or will calculate and display the p-value for you directly.
How to Show Work on the FRQ
To earn full credit for interpreting a p-value and drawing a conclusion on the AP exam, you must use a precise, two-part structure.
Part 1: Interpreting the p-value (if asked)
This script connects the definition of the p-value to the context of the problem.
- Template: "Assuming the [null hypothesis in context] is true, there is a [p-value] probability of getting a sample result as extreme or more extreme than [observed sample statistic in context] purely by chance."
Part 2: Drawing a Conclusion (always required)
This is the most common task. This script links the p-value to alpha, states the decision, and provides the conclusion in context.
Template:
"Because our p-value of [p-value] is [less than / greater than] our significance level of α = [alpha value], we [reject / fail to reject] the null hypothesis (H₀)."
"We [have / do not have] convincing statistical evidence that [alternative hypothesis in context]."
Memorize and practice these templates. They contain all the elements graders are looking for.
Practice Problems
Problem 1:
A pharmaceutical company is testing a new drug to reduce blood pressure. They conduct a clinical trial and find a p-value of 0.027 for a one-sided test of whether the drug lowers blood pressure. Using a significance level of α = 0.05, what conclusion should be drawn?
Solution:
Here, we apply the two-sentence conclusion template from the "How to Show Work on the FRQ" section.
Compare p-value to alpha and state the decision:
"Because our p-value of 0.027 is less than our significance level of α = 0.05, we reject the null hypothesis (H₀)."
State the conclusion in context:
"We have convincing statistical evidence that the new drug is effective in reducing blood pressure."
Problem 2:
A city's department of transportation claims that the mean commute time for its residents is 25 minutes. A local news organization believes this is incorrect and conducts a study. They sample 100 residents and find a sample mean commute time of 26.5 minutes. A significance test yields a p-value of 0.12.
(a) Interpret the p-value in the context of the study.
(b) Using a significance level of α = 0.10, what conclusion should the news organization make?
Solution:
(a) Interpreting the p-value: We use the interpretation template. The null hypothesis is that the mean commute time is 25 minutes. The observed sample result is a mean of 26.5 minutes.
"Assuming the true mean commute time for residents is 25 minutes, there is a 0.12 probability of getting a sample mean as extreme or more extreme than 26.5 minutes purely by chance."
(b) Drawing a conclusion: We use the two-sentence conclusion template.
Compare p-value to alpha and state the decision:
"Because our p-value of 0.12 is greater than our significance level of α = 0.10, we fail to reject the null hypothesis (H₀)."
State the conclusion in context:
"We do not have convincing statistical evidence that the true mean commute time for residents is different from 25 minutes."
Common Mistakes to Avoid
Accepting the Null Hypothesis: This is the most critical error. You MUST say "fail to reject H₀." Failing to find evidence against a claim is not the same as proving the claim is true.
Stating the p-value is the probability H₀ is true: The p-value is NOT P(H₀ is true | data). It is P(data | H₀ is true). You must always include the conditional phrase, "Assuming the null hypothesis is true..." in your interpretation.
Writing a Naked Conclusion: Simply writing "Reject H₀" or "The p-value is small" is not enough. You must always link your p-value to the significance level (α) and write your final conclusion in the context of the problem.
Forgetting "or more extreme": When interpreting a p-value, do not say it's the probability of getting exactly your sample result. The p-value includes the probability of your result and all other results that are even more unlikely.
Confusing the p-value and Alpha (α): The significance level (α) is the pre-set standard for evidence. The p-value is the calculated strength of the evidence from your specific sample. You compare the evidence (p-value) to the standard (α) to make a decision.