Understanding Scaled Scores

Students taking test
Tetra Images / Getty Images

Scaled scores are a type of exam score. They are commonly used by testing companies that administer high stakes exams, such as admissions, certification and licensure exams. Scaled scores are also used for K-12 Common Core testing and other exams that assess student skills and evaluate learning progress.

Raw Scores vs. Scaled Scores

The first step to understanding scaled scores is to learn how they differ from raw scores.

A raw score represents the number of exam questions you answer correctly. For example, if an exam has 100 questions, and you get 80 of them correct, your raw score is 80. Your percent-correct score, which is a type of raw score, is 80%, and your grade is a B-.

A scaled score is a raw score that has been adjusted and converted to a standardized scale. If your raw score is 80 (because you got 80 out of 100 questions correct), that score is adjusted and converted into a scaled score. Raw scores can be converted linearly or nonlinearly.

Scaled Score Example

The ACT is an example of an exam that uses linear transformation to convert raw scores to scaled scores. The following conversation chart shows how raw scores from each section of the ACT are transformed into scaled scores. 

Source: ACT.org
Raw Score EnglishRaw Score MathRaw Score Reading Raw Score ScienceScaled Score




The Equating Process

The scaling process creates a base scale that serves as a reference for another process known as equating. The equating process is necessary to account for differences between multiple versions of the same test.

Although test makers try to keep the difficulty level of a test the same from one version to the next, differences are inevitable.

Equating allows the test maker to statistically adjust scores so that the average performance on version one of the test is equal to average performance on version two of the test, version three of the test and so on.

After undergoing both scaling and equating, scaled scores should be interchangeable and easily comparable no matter which version of the test was taken. 

Equating Example

Let's look at an example to see how the equating process can impact scaled scores on standardized tests. Imagine that say you and a friend are taking the SAT. You will both be taking the exam at the same test center, but you will be taking the test in January, and your friend will be taking the test in February. You have different testing dates, and there is no guarantee that you will both take the same version of the SAT. You may see one form of the test, while your friend sees another. Although both tests have similar content, the questions are not exactly the same.

After taking the SAT, you and your friend get together and compare your results. You both got a raw score of 50 on the math section, but your scaled score is 710 and your friend's scaled score is 700. Your pal wonders what happened since both of you got the same number of questions correct.

But the explanation is pretty simple; you each took a different version of the test, and your version was more difficult than his. To get the same scaled score on the SAT, he would have needed to answer more questions correctly than you.

Test makers that use an equating process use a different formula to create a unique scale for each version of the exam. This means that there is no one raw-to-scale-score conversion chart that can be used for every version of the exam. That is why, in our previous example, a raw score of 50 was converted into 710 on one day and 700 on another day. Keep this in mind as you are taking practice tests and using conversion charts to transform your raw score into a scaled score.

Purpose of Scaled Scores

Raw scores are definitely easier to calculate than scaled scores.

But testing companies want to make sure that test scores can be fairly and accurately compared even if test takers take different versions, or forms, of the test on different dates. Scaled scores allow for accurate comparisons and ensure that people who took a more difficult test are not penalized, and people who took a less difficult test are not given an unfair advantage.