So you just walked out of the testing center, or maybe you're sitting at home months later wondering why your score wasn't a 5. Honestly, the 2024 AP Stats FRQ answers felt like a trap for anyone who relies too much on their calculator and not enough on their brain. It wasn't that the math was hard. It really wasn't. But the College Board went heavy on the "explain why" part this year, and that’s where things usually go south for people.
If you looked at Question 1 and thought, "Easy two-proportion z-test," you were right. But if you didn't check your conditions with the specific numbers from the problem, you probably lost points. That's the vibe of this entire set of questions. They wanted to see if you actually understood the data or if you were just a robot hitting buttons.
The Online Fitness Class Debacle (Question 1)
The first question was basically a classic inference problem. You had two groups: younger members (18-55) and older members (56+). The manager wanted to know if they had different opinions on taking online classes.
Most students nailed the hypotheses. You set the null as $p_1 = p_2$ and the alternative as $p_1
eq p_2$. Easy. But here’s the kicker: the scoring guidelines are brutal about how you define your parameters. If you just said "$p_1$ is the first group," you might have gotten a "P" for partially correct. You had to say $p_1$ is the true proportion of all exercise center members in that age group who'd be interested. "True" or "population" is the magic word there.
When it came to the conditions, the 10% rule was a big deal. You had to explicitly state that 170 is less than 10% of all younger members and 230 is less than 10% of all older members. Just writing "n < 10%" isn't enough. They want the context. The p-value came out to be around 0.35, which is huge. Since that's way higher than 0.05, you fail to reject the null. Basically, we don't have enough evidence to say their opinions are different.
Why the Bottle Fundraiser Graph Was Tricky (Question 2)
Question 2 was all about segmented bar graphs and mosaic plots. At first glance, it looks like middle school math. You just have to fill in the bars, right? Kinda.
The middle school bar had to be divided into three equal areas because the problem said the proportions were equal. But the real headache was part (c) with the mosaic plot. A mosaic plot isn't just a regular bar graph; the width of the bars represents the sample size.
High School A had a higher proportion of large bottles (0.7 vs 0.6), but High School B sold way more bottles overall. So, even though High School B's proportion was lower, the total number of large bottles they sold was higher because their bar was so much wider. If you didn't mention the width of the bars or the total sample size, you definitely left points on the table.
The Miles Per Gallon Confusion (Question 3)
This was an experimental design question. James and his Model D car. The biggest mistake people made here was confusing an observational study with an experiment.
James was just driving his own car and recording data. There was no random assignment of treatments to experimental units. Because of that, you can't conclude cause and effect. You also can't generalize the results to all cars or even all Model D cars because James only used his car. It's a sample size of one car, people!
Red Crystals and the Geometric Mean (Question 4)
Question 4 brought out the geometric distribution, which always makes people a little nervous. You had these red crystals, and you were looking for the probability of finding the first one on a certain trial.
🔗 Read more: Worcester Crime and Safety: What the Latest Shooting in Worcester MA Today Really Means
The mean of a geometric distribution is $1/p$. If the probability of finding a crystal is 0.2, the expected number of trials to find the first one is 5. Simple enough. But then they threw in the standard deviation and the shape of the distribution. Remember: geometric distributions are always skewed to the right. Always. If you said it was normal, the graders definitely sighed.
Baseball Cards and the Chi-Square Trap (Question 5)
Michelle and her baseball cards. This was a Chi-square test for independence. You had a table with categories, and you had to see if the type of card was independent of the year it was produced.
The most common error in 2024 ap stats frq answers for this one was the "Expected Counts" condition. You can't just say "the conditions are met." You have to actually list the expected counts and show that they are all at least 5. If you didn't write down the actual numbers (like 7.2, 12.4, etc.), you probably didn't get full credit.
Question 6: The Skewness Monster
The Investigative Task is always the "boss battle" of the AP Stats exam. This year, it was about whistle prices and something called Pearson’s coefficient of skewness.
They gave you a formula you’d never seen before:
$$Sk = \frac{3(\bar{x} - \text{median})}{s}$$
The goal was to see if you could use this new tool to determine if the data was skewed. You had to calculate the mean and median from a dotplot, plug them in, and interpret the result. Then, you had to discuss the normality condition for a t-interval.
Most people got tripped up because the sample size was small ($n=20$). Since 20 is less than 30, the Central Limit Theorem doesn't kick in. You have to look at the sample data. If the data shows strong skewness or outliers (which it did), you can't assume the sampling distribution of the mean is approximately normal. You had to be very specific about why the t-procedure wasn't appropriate here.
📖 Related: The Luigi Mangione Trial: What Really Happened with the New York City Shooter
How to Actually Use These Answers
If you're prepping for next year, don't just memorize these answers. Understand the patterns. The College Board loves:
- Context: Never say "the mean." Say "the mean price of the whistles."
- Conditions: Don't just list them. Show the math for $np \geq 10$ or $n < 10%$.
- Comparison: If they ask you to compare two distributions, use comparison words (higher, lower, more skewed) and address Shape, Outliers, Center, and Spread (SOCS).
Your Next Steps
Stop stressing about the exact decimal points. The AP readers care more about your logic than whether you rounded to two or three places. If you want to improve, go back to the 2024 FRQs on the College Board website and try to grade your own work using the official scoring guidelines. Pay attention to the "Notes" section in the guidelines—that's where they reveal the tiny mistakes that turn an "E" into a "P."
Once you've done that, pick a random FRQ from 2022 or 2023 and see if you can apply the same "context-heavy" approach. The more you practice writing like a statistician rather than a math student, the better you'll do.