SAT Tip: When to Guess

Many test-prep companies think they know a method that can beat the SAT scoring system Photograph by Halfdark/Getty Images

This tip on improving your SAT score was provided byShaan Patel at Veritas Prep.

Have you been told—by a friend, parent, or teacher—that it’s better to leave questions blank on the SAT than to guess, or that leaving questions blank doesn’t hurt your SAT score? If so, your adviser is oversimplifying the benefits of leaving questions blank on the SAT.

Let’s recall how the SAT is scored. The SAT raw score is calculated by:

•  Adding one point for each correct answer.

•  Subtracting ¼ point for an incorrect answer.

•  Neither adding nor subtracting for an answer left blank.

If your goal is to rack up as many raw score points as possible, it’s better not to answer a question than to answer it incorrectly. Intuitively, you may think that you are better off leaving the answer blank because you are likely to answer a question incorrectly if you guess. This is the basis of the advice from your friend, parent, or teacher. But there is more to the story.

Big test-prep companies think they know the rest of the story. Many of them offer a method that they claim “beats” the SAT scoring system.

Standard Guessing Strategy

If you can eliminate at least one of the five possible answer choices, you should randomly guess from the remaining four answer choices. If you can’t eliminate a single answer choice, you should leave the question blank.

The Reasoning

Let’s say Student A eliminates one incorrect answer choice on each of 12 SAT questions. He’s not confident enough, however, to guess by eliminating just one answer choice, so he decides to leave all 12 questions blank. His net raw score for those 12 questions would be zero.  

Now, let’s say Student B also eliminates one incorrect answer choice on each of 12 SAT questions. She randomly guesses, however, from the remaining four choices. Probability predicts that she has a one-in-four chance (25 percent) of guessing correctly. Therefore, of the 12 questions, she would guess three correctly and nine incorrectly. She would gain three points for her correct answers, and lose 2¼ for her incorrect answers, for a net raw score of +3⁄4.

The guessing strategy that Student B used is more effective, in theory. Although both Student A and Student B could eliminate only one answer choice on 12 SAT questions, Student B would be able to achieve a higher raw score because she used the guessing strategy. There is a serious flaw in this strategy, however: The method assumes that you are guessing randomly. Randomness is a difficult concept to grasp. Random guessing means that if you are presented with four answer choices, you will pick one answer choice without any prejudice, partiality, or predisposition. But that isn’t how any student “guesses” on the SAT.

Your guess is always biased in some way, because you have an inclination to one answer choice and an aversion to others. What makes matters even worse is that the College Board is really, really, really good at making wrong answer choices look appealing to unsure students, especially on difficult SAT questions. If you remove randomness from the equation, your chance of guessing the correct answer actually becomes less than 25 percent.

Veritas Prep Guessing Method

If you can eliminate at least two answer choices on a question, you should guess with as little bias as possible from the remaining answer choices. If you cannot eliminate at least two incorrect answer choices, you should leave the question blank.

The Reasoning

Let’s say that Student C eliminates two incorrect answers on each of 12 SAT questions. He then attempts to guess, with as little prejudice as possible, from the remaining three answer choices on each question. Probability predicts that if he randomly guesses from the three answer choices on each of the 12 questions, he would get four correct and eight incorrect. But because eliminating bias completely from guessing is impossible, let’s say Student C answers only three correctly and nine incorrectly. He would gain 3 points for his correct answers and 2¼ points for his incorrect answers, for a net raw score of +3⁄4. Notice that Student C’s net raw score is the same as Student B’s. But Student B’s guessing strategy is impossible, because there is no such thing as random guessing. To be sure, Student C has to eliminate two incorrect answer choices on each of the 12 questions, not just one.

Eliminate bias from guessing as much as you can. If you think that the answer to a particular question cannot be (B) because you have filled in (B) for the last three questions, you are not eliminating bias. If you think that the answer to a particular math question is (C) because it is easily derivable from the numbers in the question, you are not eliminating bias. To eliminate bias, let your pencil land on a random letter and choose that answer.

Note: For the Student-Produced Response Math questions, there is no penalty for guessing. You should answer every question in this section of the test.

Shaan Patel is the director of SAT programs at Veritas Prep, the author of McGraw-Hill’s best-selling book SAT 2400 in Just 7 Steps, and the owner of a perfect SAT score.

For more SAT advice from Veritas Prep, watch ”Perfect Scoring Student Outlines the Different Sections of the SAT”

Before it's here, it's on the Bloomberg Terminal.