Tips for Taking the Exam for AP Statistics

Tips for Multiple-Choice Questions

There are whole industries dedicated to teaching you how to take a test. A less expensive alternative for multiple-choice questions, as indicated in Section I, is to Google "test-taking tips," and check out the multiple-choice questions of some of the listings. In reality, no amount of test-taking strategy will replace knowledge of the subject. If you are on top of the subject, you'll most likely do well even if you haven't paid $500 for a test-prep course. The following tips, when combined with your statistics knowledge, should help you do well.

1. Read the question carefully before beginning. A lot of mistakes get made because students don't completely understand the question before trying to answer it. The result is that they will often answer a different question than they were asked.
2. Try to answer the question before you look at the answers. Looking at the choices and trying to figure out which one works best is not a good strategy. You run the risk of being led astray by an incorrect answer. Instead, try to answer the question first, as if there was just a blank for the answer and no choices.
3. Understand that the incorrect answers (which are called distractors) are designed to appear reasonable. Watch out for words like never and always in answer choices. These frequently indicate distractors. Don't get suckered into choosing an answer just because it sounds good! The question designers try to make all the logical mistakes you might make and the answers they come up with become the distractors. For example, suppose you are asked for the median of the five numbers 3, 4, 6, 7, and 15. The correct answer is 6 (the middle score in the ordered list). But suppose you misread the question and calculated the mean instead. You'd get 7 and, be assured, 7 will appear as one of the distractors.
4. Drawing a picture can often help visualize the situation described in the problem. Sometimes, relationships become clearer when a picture is used to display them. For example, using Venn diagrams can often help you "see" the nature of a probability problem. Another example would be using a graph or a scatterplot of some given data as part of doing a regression analysis.
5. Leave an answer blank if you have no clue what the answer is. The "guess penalty" is designed to yield an average score of 0 for guessed questions (that is, if you guessed at every questions, you'd get, on average, 0—you're a statistics student, figure out how that works!). On the other hand, if you can definitely eliminate at least one choice, it is to your advantage in the long run to guess. You don't have to answer every question to get a top score.
6. Double check that you have (a) answered the question you are working on, especially if you've left some questions blank (it's horrible to realize at some point that all of your responses are one question off!) and (b) that you have filled in the correct bubble for your answer. If you need to make changes, make sure you erase completely and neatly.

Tips for Free-Response Questions

There are many helpful strategies for maximizing your performance on free-response questions, but actually doing so is a learned skill. Students are often too brief, too sloppy, or too willing to assume the reader will fill in the blanks for them. You have to know the material to do well on the free-response, but knowing the material alone is not sufficient—you must also demonstrate to the reader that you know the material. Many of the following tips will help you do just that.

1. Read all parts of a question first before beginning. There's been a trend in recent years to have more and more sub-parts to each question (a, b, c,…). The sub-parts are usually related and later parts may rely on earlier parts. Students often make the mistake of answering, say, part (c) as part of their answer to part (a). Understanding the whole question first can help you answer each part correctly.
2. WRITE LEGIBLY!!! I was a table leader for the AP Statistics Exam for seven years and nothing drove me, or other readers, crazier than trying to decipher illegible scribbling. This may sound silly to you, but you'd be amazed at just how badly some students write! It doesn't need to look like it was typewritten, but a person with normal eyesight ought to be able to read the words you've written with minimal effort.
3. Use good English, write complete sentences, and organize your solutions. You must make it easy for the reader to follow your line of reasoning. This will make the reader happy and it's in your self-interest to make the reader (very) happy. The reader wants you to do well, but has only a limited amount of time to dedicate to figuring out what you mean. Don't expect the reader to fill in the blanks for you and make inferences about your intent—it doesn't work that way. Also, answer questions completely but don't ramble. While some nonsense ramblings may not hurt you as long as the correct answer is there, you will be docked if you say something statistically inaccurate or something that contradicts an otherwise correct answer. Quit while you are ahead. Remember that the amount of space provided for a given question does not necessarily mean that you should fill the space. Answers should be complete but concise. Don't fill space just because it's there. When you've completely answered a question, move on.
4. Answers alone (sometimes called "naked" answers) may receive some credit but usually not much. If the correct answer is "I'm 95% confident that the true proportion of voters who favor legalizing statistics is between 75% and 95%" and your answer is (0.75, 0.95), you simply won't get full credit. Same thing when units or measurement are required. If the correct answer is 231 feet and you just say 231, you most likely will not receive full credit.
5. Answers, and this is important, must be in context. A conclusion to an inference problem that says, "Reject the null hypothesis" is simply not enough. A conclusion in context would be something like, "At the 0.05 level of significance, we reject the null hypothesis and conclude that there is good evidence that a majority of people favor legalizing statistics."
6. Make sure you answer the question you are being asked. Brilliant answers to questions no one asked will receive no credit. (Seriously, this is very common—some students think they will get credit if they show that they know something, even if it's not what they should know at the time.) Won't work. And don't make the reader hunt for your final answer. Highlight it in some way.
7. Simplify algebraic or numeric expressions for final answers. You may still earn credit for an unsimplified answer but you'll make the reader work to figure out that your answer is equivalent to what is written in the rubric. That will make the reader unhappy, and, as mentioned earlier, a happy reader is in your best interest.
8. If you write a formula as part of your solution, use numbers from the question. No credit is given for simply writing a formula from a textbook (after all, you are given a formula sheet as part of the exam; you won't get credit for simply copying one of them onto your test page). The reader wants to know if you know how to use the formula in the current problem.
9. If you are using your calculator to do a problem, round final answers to two or three decimal places unless specifically directed otherwise. Don't round off at each step of the problem as that creates a cumulative rounding error and can affect the accuracy of your final answer. Also, avoid writing calculator syntax as part of your solution. The readers are instructed to ignore things like normalcdf, 1PropZTest, etc. This is called "calculatorspeak" and should not appear on your exam
10. Try to answer all parts of every question—you can't get any credit for a blank answer. On the other hand, you can't snow the readers—your response must be reasonable and responsive to the question. Never provide two solutions to a question and expect the reader to pick the better one. In fact, readers have been instructed to pick the worse one. Cross out clearly anything you've written that you don't want the reader to look at.
11. You don't necessarily need to answer a question in paragraph form. A bulleted list or algebraic demonstration may work well if you are comfortable doing it that way.
12. Understand that Question #6, the investigative task, may contain questions about material you've never studied. The goal of such a question is to see how well you think statistically in a situation for which you have no rote answer. Unlike every other question on the test, you really don't need to worry about preparing for this question above normal test preparation and being sure that you understand as much of the material in the course as possible.

Specific Statistics Content Tips

The following set of tips are things that are most worth remembering about specific content issues in statistics. These are things that have been consistent over the years of the reading. This list is not exhaustive! The examples, exercises, and solutions that appear in this book are illustrative of the manner in which you are expected to answer free-response problems, but this list is just a sampling of some of the most important things you need to remember.

1. When asked to describe a one-variable data set, always discuss shape, center, and spread.
2. Understand how skewness can be used to differentiate between the mean and the median.
3. Know how transformations of a data set affect summary statistics.
4. Be careful when using "normal" as an adjective. Normal refers to a specific distribution, not the general shape of a graph of a data set. It's better to use approximately normal, mound-shaped, bell-shaped, etc., instead. You will be docked for saying something like, "The shape of the data set is normal."
5. Remember that a correlation does not necessarily imply a real association between two variables (correlation is not causation). Conversely, the absence of a strong correlation does not mean there is no relationship (it might not be linear).
6. Be able to use a residual plot to help determine if a linear model for a data set is appropriate. Be able to explain your reasoning.
7. Be able to interpret, in context, the slope and y-intercept of a least-squares regression line.
8. Be able to read computer regression output.
9. Know the definition of a simple random sample (SRS).
10. Be able to design an experiment using a completely randomized design. Understand that an experiment that utilizes blocking cannot, by definition, be a completely randomized design.
11. Know the difference between the purposes of randomization and blocking.
12. Know what blinding and confounding variables are.
13. Know how to create a simulation for a probability problem.
14. Be clear on the distinction between independent events and mutually exclusive events (and why mutually exclusive events can't be independent).
15. Be able to find the mean and standard deviation of a discrete random variable.
16. Recognize binomial and geometric situations.
17. Never forget that hypotheses are always about parameters, never about statistics.
18. Any inference procedure involves four steps. Know what they are and that they must always be there. And never forget that your conclusion in context (Step 4) must be linked to Step 3 in some way.
19. When doing inference problems, remember that you must show that the conditions necessary to do the procedure you are doing are present. It is not sufficient to simply declare them present.
20. Be clear on the concepts of Type I and Type II errors and the power of a test.
21. If you are required to construct a confidence interval, remember that there are three things you must do to receive full credit: justify that the conditions necessary to construct the interval are present; construct the interval; and interpret the interval in context. You'll need to remember this; often, the only instruction you will see is to construct the interval.
22. If you include graphs as part of your solution, be sure that axes are labeled and that scales are clearly indicated. This is part of communication.