The Big Picture
Welcome to inference for means! In the last unit, you mastered making claims about proportions (like the percent of people who support a candidate). Now, we shift our focus to quantitative data—things we can measure, like height, weight, time, or GPA. The core ideas of confidence intervals and significance tests remain the same, but our tools will change.
Think of it like being a quality control manager at a factory that makes 16-ounce bags of chips. You can't weigh every single bag, so you take a random sample. Your sample might have an average weight of 15.9 ounces. Is that close enough to 16, or is your machine systematically under-filling the bags? This unit gives you the statistical tools to answer that question. The big new idea is the t-distribution, a special tool we use when we have to estimate the population standard deviation from our sample, which is almost always the case in real-world scenarios.
Key Questions
How can we create a range of plausible values for a population's true average when we only have data from one sample?
How do we formally test a claim about a population mean (e.g., "The average battery life is at least 10 hours")?
What is the difference between a one-sample test, a two-sample test, and a paired test, and how do I know which one to use?
Why do we use a t-distribution instead of the familiar normal distribution for means, and what are "degrees of freedom"?
Your Learning Path
1. Inference for a Single Population Mean
Topic 7.1 - 7.5: Confidence Intervals and Significance Tests for One Mean
You'll begin by learning the fundamental building blocks for inference with a single quantitative variable. You will learn why we need the t-distribution when we don't know the population standard deviation. You'll then apply the familiar four-step process to both construct and interpret a confidence interval for a population mean, and then to set up and carry out a full significance test for a claim about a population mean.
2. Inference for Comparing Two Population Means
Topic 7.6 - 7.9: Confidence Intervals and Significance Tests for the Difference of Two Means
This section expands our skills to situations where we want to compare the means of two independent groups. For example, is there a significant difference in the average GPA of students who play sports versus those who don't? You'll learn how to build confidence intervals for the difference between two means and how to perform a significance test to see if a measured difference is statistically significant or just due to random chance.
3. Choosing the Right Tool for the Job
Topic 7.10: Selecting, Implementing, and Communicating Inference Procedures
This topic is the capstone of the unit. You'll practice analyzing a problem description and selecting the correct inference procedure from all the ones you've learned so far (for both proportions and means). This is a critical skill for the AP exam, as you'll need to distinguish between one-sample, two-sample, and paired data scenarios and justify your choice.
How to Succeed in This Unit
Conditions are Critical: For means, the "Normal/Large Sample" condition is key. You must explicitly check it. State one of the following: the population is stated to be normal, the sample size is large (n ≥ 30) so the Central Limit Theorem applies, or a graph of the sample data (like a dotplot or boxplot) shows no strong skewness or outliers. Don't just say "the sample is normal."
Master the t-Distribution: Remember the key difference: use z for proportions, use t for means (unless the population standard deviation, σ, is somehow known, which is rare). When you perform a t-procedure, you must report the degrees of freedom (df). For a one-sample t-test, df = n - 1. For a two-sample t-test, use your calculator's value or the conservative smaller of n₁-1 and n₂-1.
Don't Confuse Two-Sample and Paired Tests: This is a very common mistake. If the data consists of two measurements on the same subject (e.g., before and after a treatment) or on two matched subjects, it is paired data. To analyze it, you first find the difference for each pair and then perform a one-sample t-test on the differences. If the data comes from two completely independent groups, you use a two-sample t-test.
Communicate Your Conclusion in Context: Never end your conclusion with just "I reject the null hypothesis." Always link your decision back to the original question. For example, "Because our p-value of 0.02 is less than α = 0.05, we reject the null hypothesis. We have convincing evidence that the true mean battery life of this brand of phone is greater than 10 hours."