The Big Picture
Welcome to the world of statistical inference! Until now, you've focused on describing data from a sample. In this unit, you'll learn how to use that sample data to make a reasonable guess—or test a specific claim—about the entire population. We're moving from "what does our data say?" to "what can our data tell us about the world at large?"
Think of it like taste-testing a giant pot of soup. You can't eat the whole pot, so you take one spoonful (your sample). Based on that spoonful, you make an inference about the entire pot (the population). Is it salty enough? You might say, "I'm pretty sure the whole pot is well-seasoned." This unit gives you the statistical tools to say exactly how sure you are and to formally test a claim like, "The recipe's claim that less than 10% of the beans are undercooked is wrong." We'll focus exclusively on categorical data—situations with "yes/no" or "success/failure" outcomes, which we measure with proportions.
Key Questions
How can I use a proportion from a single sample to build a range of plausible values for the true proportion of an entire population?
What does it mean to be "95% confident" in a statistical finding, and how do I communicate that level of uncertainty?
How can I use probability to determine if a sample result is so unusual that it provides convincing evidence against a claim?
When comparing two groups (like a treatment and a control group), how do I decide if the observed difference between them is real or just due to random chance?
Your Learning Path
1. Inference for a Single Proportion: Estimation
Topic 6.1 - 6.3: From Sample to Population Estimate
You'll begin by exploring the foundation of all inference: the sampling distribution. You'll see why the Normal model is so crucial for making predictions about a sample proportion. With that foundation, you will master the construction and interpretation of a confidence interval for a single population proportion. This is the primary tool for estimating an unknown population value and communicating our uncertainty using a margin of error.
2. Inference for a Single Proportion: Hypothesis Testing
Topic 6.4 - 6.7: Making a Decision About a Claim
Here, you'll shift from estimating a value to testing a claim. You will learn the formal, four-step process for a significance test. This involves stating hypotheses, checking conditions, calculating a test statistic and a p-value, and making a conclusion in context. You'll gain a deep understanding of the p-value—one of the most important and often misinterpreted concepts in statistics—and learn about the types of errors (Type I and Type II) we can make when we perform a test.
3. Comparing Two Proportions
Topic 6.8 - 6.11: Is There a Real Difference?
You will now apply everything you've learned to one of the most common statistical scenarios: comparing two different populations or two treatment groups. You will learn how to construct and interpret confidence intervals for the difference between two proportions. You'll also learn how to carry out a full significance test to determine if there is a statistically significant difference between the two groups, a skill essential for interpreting scientific studies and A/B testing.
How to Succeed in This Unit
Conditions are King: For every confidence interval or significance test you perform, you must state and check the required conditions (Random, 10% Rule, Large Counts). On the AP Exam, failing to check conditions in the context of the problem is a guaranteed way to lose points. Don't just name them; show that they are met with numbers from the problem.
Master the Four-Step Process: Structure every inference problem using a consistent framework: State (hypotheses and parameters), Plan (name the procedure and check conditions), Do (show calculations), and Conclude (interpret the result in context). This organization makes your reasoning clear and ensures you don't miss any required elements.
Interpret, Don't Just State: The number is not the answer. The answer is what the number means. Memorize and practice using these sentence frames:
Confidence Interval: "We are [C]% confident that the interval from ___ to ___ captures the true proportion of [context]."
P-Value: "Assuming the null hypothesis is true (in context), there is a [p-value] probability of getting a sample proportion as or more extreme than the one we observed."
Mind Your p's and p̂'s: Notation is critical. Use p for the population proportion (the parameter you're making a claim about) and p̂ ("p-hat") for the sample proportion (the statistic you calculated from your data). Mixing these up in your hypotheses or formulas is a common error that shows a misunderstanding of the core concepts.