For years prior to the release of the digital SAT, we had conversations with College Board employees about their plans to create a shorter exam. Of course, students would prefer the SAT to be as brief as possible. However, a test that is too short won’t provide enough data. The problem for test-makers is to design a test that has fewer questions without compromising the validity of its scoring.
A shorter SAT required weighted questions
The digital SAT achieved the necessary balance of brevity and validity with its adaptive structure and question weighting. The College Board spent a long time choosing their ideal adaptive test format, which adjusts the difficulty of questions based on each student’s performance and allows the digital SAT to focus more on the questions that provide the most useful data for each student.
The SAT’s adaptive format is well documented and relatively easy to understand, but the test’s system of question weighting is more obscured and complex. Currently, you won’t find any official documentation on the SAT’s question weighting (though the College Board is planning to publish an overview of these specifications eventually), but we have had private conversations with College Board’s psychometricians about these systems, and they are verifiable with thorough analysis of test data.
Put simply, the digital SAT’s questions are not all worth the same.
Some SAT questions are worth more than others are
Throughout the SAT, some questions have a greater impact on scores than others do. The varied impact of different questions is called weighting. These differences in question weights affect your scaled scores, and they are also factors in determining whether you get into the Higher or Lower modules on the test.
To highlight this feature of the SAT’s scoring, consider the following example. In the official practice SAT test #1, if you only miss question #25 in Module 2 of the Reading & Writing section, you will get a 780. In the same section, if you only miss question #23, you will get a 750. Why does #23 have more than twice as much impact on scores than #25 does? The two questions are very similar, testing the same skills, but the key difference is that #23 is significantly easier.
Question weighting should be a major consideration in every student’s testing strategy. To improve your SAT scores, you should focus on not only answering more questions correctly but also answering the most impactful questions correctly.
Which questions have the most impact on your scores? Our test analysis shows that the questions with the greatest weighting are easy questions and Student-Produced Responses.
Easy questions are important
Throughout the SAT, easy questions have a greater impact than harder questions do. This scoring system may seem counterintuitive, as the most challenging questions on the test are worth less than the easiest questions are, but it is based on a statistical analysis system known as Item Response Theory, which adjusts the weighting of questions based on how well they indicate test-takers’ abilities. Put simply, high-scoring students usually get easy questions right, so missing an easy question means you’re less likely to get a high score.
Give due respect to the easy questions. In the Math section, read carefully to ensure you are answering the question being asked, and avoid careless mistakes such as missing a negative sign or miscalculating a decimal value.
In the Reading & Writing section, the differences between answers can be subtle and some answers are almost correct, so it’s important to read through all of the choices, even if you quickly find one that looks right. If you rush and lose your focus because questions seem too easy, you might make a minor error.
Missing a single easy question can have as much of an impact on your score as two or three of the hardest questions would.
Student-Produced Responses are tough but critical
In the Math section, the Student-Produced Responses (those questions that require you to enter a numerical answer) are also worth more. Because you can’t simply guess on these questions, a correct answer is a sure sign that you understand the material, which is why they are weighted more. These questions can be challenging, especially because they do not offer answer choices, but you should never skip them.
The statistical engine under the hood: Item Response Theory
Item Response Theory (IRT) is used to analyze question data and determine the relative impact of questions. IRT is also known as latent trait theory, strong true score theory, or modern mental test theory. It’s a system of psychometric analysis that’s used to design, analyze, and score tests.
IRT models predict the likelihood of students to correctly answer test questions. An important parameter in IRT models is item discrimination, the rate at which the probability of getting a correct answer changes at different ability levels.
On the SAT, questions with higher discrimination have a bigger impact on scores. Easy questions and Student-Produced Responses tend to have high discrimination, which is why they have a larger impact on scores.
Focus on what matters most
The digital SAT is a necessarily complex test, but you don’t need to understand all the nuances of its adaptive structure or IRT model to perform your best on the test. You can optimize your testing strategies by recognizing the importance of easy questions and Student-Produced Responses.
For the former, read the questions carefully, make sure you’re giving the proper answer that the question is asking for, and quickly check your work. For the latter, don’t skip them just because they seem more challenging, and practice them extensively if they cause you anxiety. With this understanding of the SAT’s weighted question system, you can easily improve your score by prioritizing the questions with the greatest impact.
The post SAT Question Weighting: What Matters Most on the Test appeared first on Summit Educational Group.