Note: this old article has been superseded by a new blog post that reflects the revisions made to the TOEFL in 2023!
TOEFL Raw to Scaled Scores: An Explanation
In the TOEFL reading section you answer 30 questions, worth up to 33 points. This is your raw reading score.
In the TOEFL listening section you answer 28 questions, worth up to 28 points. This is your raw listening score.
For your final score report these are each converted to scores out of 30. These are called your scaled scores.
Note that the unofficial reading and listening scores displayed on your screen at the end of the TOEFL test are also scaled.
Note also that you might be given an extra set of ten additional reading questions, or a set of 11 additional listening questions. You will never know which set is “extra.” These are not scores.
The “Official” Reading Conversion
Here is the raw to scaled conversion chart from the latest Official Guide to the TOEFL (August, 2020):
You can see that the raw scores convert to a range of scaled scores. For instance, if you get 27 points on the actual test, you might get a final score of 25, 26, 27 or 28 points. This reflects the fact that the conversion is different for every single test! Based on this chart we might assume that you can always get a perfect score with only 32 points… but don’t take this chart as a guarantee. It is just a general guideline.
The “Official” Listening Conversion
Here is the raw to scaled conversion chart from the latest Official Guide to the TOEFL (August, 2020):
Once again, the raw scores convert only to a range of scaled scores. Every test is different.
The Actual Raw to Scaled Conversion Method is a Secret
The actual raw to scaled conversion method is a secret. As indicated, every test has a different conversion. This means that sometimes you need to answer every single question correctly to get a perfect score, but sometimes you can make a few mistakes and still get a perfect score. The conversion is based on the difficulty of the specific test. Here’s what ETS says:
“…despite the best efforts of test designers, no two test forms can ever be exactly alike. Each form is composed of different questions, which means that each test form differs slightly from other test forms in its level of difficulty. Therefore, number-correct (raw) scores are bound to a specific test form and are not directly comparable across forms.” (source)
The difficulty of a given test is determined through a process called equating. In practice, this means that each test contains questions that have been “pre-tested” in the past. I assume these questions are contained in the extra sets mentioned above.
A Rough Raw-to-Scaled Reading Chart without Ranges
If you want to score a practice set and the creators have not provided a conversion chart, you can make your own “best guess” chart using basic math. Here’s the easy formula: (RAW/33) * 30 = SCALED. Round the fractions. If your practice test contains extra questions adjust as needed. Here’s what I came up with:
A Rough Raw-to-Scaled Listening Chart without Ranges
Likewise, you can make a basic conversion chart for the listening section of the TOEFL with the following formula: (RAW/28) * 30 = SCALED. That chart looks like this:
How about the Speaking and Writing Sections?
Oh, who the f–k knows? I’ll write an article about them in the future, but now that the e-rater and SpeechRater are used it is impossible to know exactly how the questions are graded. We know how about the raw scores from human graders, but the AI technology is mysterious. We don’t even know how the human and AI are weighed. Just note that the speaking and writing scores are adjusted up and down depending on the difficulty of a given test. I think.
Sign up for express essay evaluation today!
Submit your practice essays for evaluation by the author of this article. Get feedback on grammar, structure, vocabulary and more. Learn how to score better on the TOEFL. Feedback in 48 hours.