We know that studying for your AP exams can be stressful, but Fiveable has your back! We have created a study plan that will help you crush your statistics exam. We will continue to update this guide with more information about the 2022 exams, as well as helpful resources to help you score that 5.
👉
Try using a
study timer like the one in Fiveable rooms to maximize your efficiency when preparing for the exam!
This year, all AP exams will cover all units and essay types. The 2022 Statistics exam format will be:
Section 1: Multiple Choice - 50% of your score
Section 2: Free Response - 50% of your score
Part A: 65 mins
1 multipart question with a focus on collecting data
1 multipart question with a focus on exploring data
1 multipart question with a focus on probability and sampling distributions
1 multipart question with a focus on inference
1 multipart question that combines 2 or more skill categories
Part B: 25 mins
Check out our study plan below to find resources and tools to prepare for your AP Statistics exam.
The exam is on paper, in school, on Thursday, May 5, 2022, at 12 PM, your local time.
First, download the AP Statistics Cram Chart PDF - a single sheet that covers everything you need to know at a high level. Take note of your strengths and weaknesses!Review every unit and question type, and focus on the areas that need the most improvement. We've put together this plan to help you study between now and May. This will cover all of the units and essay types to prepare you for your exam.Additionally, create your schedule ahead of time so that you can connect with other students instead of just studying alone! Join Hours 🤝to talk to real students just like you studying for this exam. Finally, check out our Cram replays so that you can review for the AP Stats exam with a rockstar teacher and study socially among other students.Before we begin, take some time to get organized. Remote learning can be great, but it also means you’ll need to hold yourself accountable more than usual.
Make sure you have a designated place at home to study. Somewhere you can keep all of your materials, where you can focus on learning, and where you are comfortable. Spend some time prepping the space with everything you need and you can even let others in the family know that this is your study space.
Get your notebook, textbook, prep books, or whatever other physical materials you have. Also create a space for you to keep track of review. Start a new section in your notebook to take notes or start a Google Doc to keep track of your notes. Get yourself set up!
The hardest part about studying from home is sticking to a routine. Decide on one hour every day that you can dedicate to studying. This can be any time of the day, whatever works best for you. Set a timer on your phone for that time and really try to stick to it. The routine will help you stay on track.
How will you hold yourself accountable to this study plan? You may or may not have a teacher or rules set up to help you stay on track, so you need to set some for yourself. First, set your goal. This could be studying for x number of hours or getting through a unit. Then, create a reward for yourself. If you reach your goal, then x. This will help stay focused!
There are thousands of students all over the world who are preparing for their AP exams just like you! Join
Hours 🤝
to chat, ask questions, and meet other students who are also studying for the spring exams. You can even build study groups and review material together!
Unit 1 is about creating and analyzing graphs of data. This includes both categorical and quantitative data. For categorical data, we should be able to read and create tables and bar graphs and calculate proportions/percentages. For quantitative data, we should be able to read and create dot plots, stemplots, histograms, and boxplots. We should also be able to describe the shape, center, variability (spread), and any unusual features of a distribution of quantitative data. This includes making calculations such as mean, median, range, interquartile range (IQR), and standard deviation. Our descriptions and calculations can be used to compare data from multiple groups. Finally, Unit 1 ends with describing the position of individuals within a quantitative data set, including using percentiles and z-scores. This leads us to an initial exploration of the Normal Distribution, though we will study that more in-depth in Units 4-5.
Unit 2 is about creating and analyzing graphs of data when two variables are measured about each individual in a data set. For categorical data, we should be able to read and create two-way tables or segmented bar graphs and calculate conditional percentages. These can be used to comment on the association (or lack thereof) between the two variables. For quantitative data, we should be able to read, create, and describe scatterplots, which can also be used to comment on the apparent association between two variables.
The second half of Unit 2 is then focused on linear regression, a process by which we can make predictions about one quantitative variable (a response variable) using another (an explanatory variable). We should be able to use Least-Squares Regression Lines to make these predictions, and interpret several components of the LSRLs (including slope, intercept, and other calculated values such as s or r2)
While Units 1-2 were about graphing and analyzing sets of data, Unit 3 is about examining the methods through which we can collect that data. For sample surveys, we should be able to describe various methods of selecting samples, particularly the random methods (simple random, stratified random, cluster, and systematic samples). However, not all samples are collected through a random process, and we should be prepared to discuss possible sources of bias in surveys (including via non-random selection processes).
We then turn to the differences between observational studies and experiments, and the features of a well-designed experiment. We should be able to define many common terms associated with experiments (many of which you’ve likely seen in other courses!), and compare and contrast several common experimental designs: completely randomized design, randomized block design, and matched-pairs design.
Unit 4 is where AP Statistics gets “math-y,” with lots of calculations and formulas. We are asked to calculate or interpret probabilities in a variety of settings, beginning with the understanding that probability reflects what we should expect to occur over the long run. We should be able to design and execute simulations for a given scenario - and then the calculations begin. We should be able to calculate the probability of multiple events using a variety of strategies (including Two-Way Tables, Tree Diagrams, and/or Venn Diagrams).
We should also be able to categorize different events as “mutually exclusive” or “independent,” with justification. Conditional probability [P(A | B)] plays a big role in this part of the unit. Shifting over to random variables, we should be able to calculate the mean (expected value) or standard deviation of a random variable, and combine them using similar rules to Unit 1. We conclude Unit 4 with a look at Binomial and Geometric random variables, which are two special types of variables that arise frequently in applications.
Unit 5 provides the bridge from descriptive statistics (Units 1-4) to inferential statistics (Units 6-9). After reviewing the Normal Distribution and introducing the idea of using sample statistics (like p or x) to estimate population parameters, we explore the creation of sampling distributions.
We meet the conditions for inference: random samples, large samples (for categorical variables, we need at least 10 expected successes and failures; for quantitative variables, we need n to be at least 30), and independent observations (which turns into the “10% rule” for sampling without replacement: if the sample size n is less than 10% of the population size N, we can do calculations as if we sampled with replacement).
If these conditions are met, the sampling distribution we build will be approximately Normal and all of our formulas for calculating the mean and standard deviation of sampling distributions on the formula sheet will hold. We then build sampling distributions for sample proportions/sample means and the difference of sample proportions/sample means.
Unit 6 is where we meet Confidence Intervals and Hypothesis Tests for the first time, specifically z-intervals and z-tests for population proportions. After learning “the basics” about confidence intervals (what’s a confidence level? What’s a margin of error?), we construct and interpret 1 and 2-sample z-intervals.
These intervals, built from samples, can be used to justify claims about a population. Then, after exploring the rationale behind hypothesis tests (including how to write null/alternative hypotheses and interpret a p-value in context), we run 1 and 2-sample z-tests. Finally, we meet “Errors”: both Type I (rejecting a true H0) and Type II (failing to reject a false H0), and define the “Power” of a test as the probability of correctly rejecting a false H0. This unit is often heavily tested and is well worth your time to review!
Unit 7 is an extension of Unit 6: we basically do everything again, but with t-procedures instead of z-procedures! We build Confidence Intervals and run Hypothesis Tests for a population mean or a difference of population means.
For the difference of population means, we must be able to distinguish between if we are running a 2-sample procedure or a matched-pairs procedure (in which we will use a 1-sample procedure to execute the process).
Unit 8 is where we learn about chi-square tests, which can be used when there are two or more categorical variables. We’ll learn how to select from the following tests: the chi-square test for goodness of fit (for a distribution of proportions of one categorical variable in a population), the chi-square test for independence (for associations between categorical variables within a single population), or the chi-square test for homogeneity (for comparing distributions of a categorical variable across populations or treatments).
Unit 9 will teach students how to construct confidence intervals for and perform significance tests about the slope of a population regression line when appropriate conditions are met. Surprisingly, there is variability in slope, which differs from students’ experience in previous courses. Slopes will likely vary as part of an approximately normal sampling distribution centered at the (true) slope of the population regression line relating spring length to hanging mass.
🎥 Watch these videos:
✍️ Practice: