How 47 TOEFL Listening items become a band from 1.0 to 6.0, what the 60 percent routing threshold does to your ceiling, and where half a band is won.
11 min readMonthly & Premium
Quick answer
Every Listening item is worth one raw point. The real exam gives you 47 items and scores 35 of them, using the rest to trial questions for future tests, while our practice tests score all 47. Raw converts to a scaled score out of 30, which converts to a band from 1.0 to 6.0 in half-point steps. Scoring 60 percent or more of the first module routes you to the harder second module and removes the band 4.0 ceiling.
Every item is worth exactly one point
Start here, because it changes how you should spend your attention. In the shipped Listening tests, every question carries one point. A one-sentence Choose a Response item is worth the same as the inference question buried at the end of a three-paragraph lecture on the Zeigarnik Effect. There is no bonus for difficulty.
That means the fastest way to lose marks is to spend disproportionate effort on hard items. If a lecture question is genuinely ambiguous to you, the correct economic decision is to answer it in five seconds and protect the next three short items, which are far more reliably winnable. Students who do the opposite, agonising over the lecture and arriving flustered at the reply items, routinely score below their actual comprehension level.
It also means there is nothing to gain from leaving an item blank. There is no penalty for a wrong answer. An unanswered item and a wrong answer are worth the same zero, so a guess is free upside. If you take nothing else from this lesson, take that: never submit a blank Listening item.
With 47 items in front of you, the range you are playing in here is 0 to 47 raw points. On the real exam the scored range is 0 to 35, because ETS counts 35 of the 47 and uses the others to trial questions for future tests. It does not publish which ones, so the practical advice is identical either way: answer all 47. What happens to your raw number next is where most people's mental model is wrong.
How your clicks become a band
1
Step 1: raw score
Your correct answers are counted. One point each, across both modules, for a maximum of 47.
2
Step 2: scaled score out of 30
Your raw score is converted to a scaled score on a 0 to 30 range using a conversion table. The conversion is not linear. In the middle of the range roughly two raw points buy one scaled point, but near the top the table compresses, so the last few raw points are worth less than you would expect.
3
Step 3: scaled to band
The scaled score maps onto the 2026 band scale, 1.0 to 6.0 in 0.5 steps. The thresholds are fixed: scaled 29 or above is band 6.0, 27 is 5.5, 24 is 5.0, 21 is 4.5, 18 is 4.0, 15 is 3.5, 12 is 3.0, 9 is 2.5, 6 is 2.0, 3 is 1.5, and below that is band 1.0.
4
Step 4: the adaptive cap
If your first-module performance routed you to the easier second module, your Listening band is capped at 4.0 regardless of how well you did in that easier module. This is the step almost nobody knows about, and it is the reason the first module deserves your best attention.
5
Step 5: your section band joins the overall
Your Listening band sits alongside Reading, Speaking and Writing. The overall band is the average of the section bands you actually took, rounded to the nearest 0.5. One weak section drags the average by roughly a quarter of its own gap, so a Listening band a full point below your others costs you about 0.25 overall.
Raw score to Listening band
Raw correct (of 47)
Scaled (of 30)
Band
15 to 18
9 to 11
2.5
19 to 22
12 to 14
3.0
23 to 26
15 to 17
3.5
27 to 31
18 to 20
4.0
32 to 35
21 to 23
4.5
36 to 38
25 to 26
5.0
39 to 40
27 to 28
5.5
41 or more
29 to 30
6.0
How raw correct answers out of 47 convert on our practice tests. Bands are chunky, so the marginal value of one more correct answer varies enormously depending on where you are.
Where an extra correct answer is actually worth something
Look at the band 4.0 row in that table. It spans five raw scores, from 27 to 31. If you score 27 and improve to 31, you have got four more questions right and your reported band has not moved at all. Then one more correct answer, the thirty-second, moves you to 4.5.
This is not a flaw in the scale, it is what any banded scale does. But it has a practical consequence: your improvement will feel invisible for a while and then arrive all at once. Students who track only their band get discouraged in exactly the stretch where they are actually improving. Track your raw score out of 47 in practice. It moves smoothly, it tells you the truth, and it lets you see progress that the band is hiding.
The other place the table matters is the top. Between band 5.5 and band 6.0 there is a difference of two raw answers. At that level, one careless click on a short reply item is the whole difference. This is why strong listeners should spend revision time on Choose a Response rather than on lectures. The lecture items were never the ones costing them the last band.
At the bottom of the useful range, the gap between band 3.0 and band 4.0 is eight raw answers, from 19 to 27. Eight items is a realistic month of focused work if you know which task type is leaking. That is what the task lessons in this course are for.
What each band means in CEFR terms
Band
CEFR
What a listener at this level can do
6.0
C2
Follows a fast academic lecture and a rapid two-way conversation with no measurable loss, including implied attitude and dry humour.
5.0 to 5.5
C1
Follows extended academic speech comfortably. Handles connected speech and unfamiliar topics. Occasional loss on very fast or heavily accented delivery.
4.0 to 4.5
B2
Follows the main line of a lecture and gets most detail. Loses ground on implied meaning, rapid exchanges and unfamiliar vocabulary.
3.0 to 3.5
B1
Gets the topic and clearly stated facts. Regularly misses purpose, attitude and anything not said in so many words.
2.0 to 2.5
A2
Understands short, slow, clearly structured speech on familiar subjects. Extended lectures are largely out of reach.
1.0 to 1.5
A1
Catches isolated familiar words and very short simple statements.
The 2026 band scale is aligned to CEFR. This is the mapping used across our score reporting.
Reading your own practice results properly
A single band number tells you almost nothing about what to do next. Two students at band 3.5 can need completely opposite training. Break your result down by task type before you draw any conclusion.
If you lose points mostly on Choose a Response, your problem is speed of intent recognition, not comprehension. You are probably understanding every word and still picking the option that repeats vocabulary from the clip. Go to the Choose a Response lesson and the traps lesson.
If you lose points mostly on conversations, your problem is holding two speakers apart. You know what was said and not who said it or who agreed to do it. The conversation lesson and the note-taking lesson are the fix, in that order.
If you lose points mostly on announcements, your problem is detail capture under time pressure. Announcements are the most predictable material in the section, so this is the cheapest gap in the whole test to close.
If you lose points mostly on academic talks, separate the fact questions from the inference and next-topic questions. Losing facts means your notes are failing. Losing inference means you are transcribing instead of listening. Those need different practice.
One more diagnostic worth running: count your first-module percentage separately from your overall. If you are above 60 percent overall but below it in the first module, you have a warm-up problem, and warm-up problems are the fastest thing in this entire course to fix.
Check you have the scoring model right
Three questions on how the numbers actually behave. Getting these wrong leads to real strategy mistakes on test day.
You are 20 minutes into Listening and one lecture question is genuinely unclear. Three short reply items are coming next. What is the correct decision?
Every item is worth one raw point regardless of difficulty, and there is no penalty for a wrong guess. A quick guess on the hard item costs nothing you were likely to get anyway, and the three short items that follow are far more reliably winnable.Difficulty does not change an item's value. Every Listening item carries one point, so a hard lecture question and a one-line reply item are worth exactly the same.Blank and wrong score the same zero. Leaving it blank throws away a free chance at a point and does not protect you from anything.This is the most common version of the mistake. Trading three likely points for one unlikely point is a losing trade every time.
Two students each answer 30 of 47 items correctly. Student A scores 19 of 30 in the first module; Student B scores 16 of 30. Why can their reported bands differ?
Routing happens on first-module performance. At or above 60 percent you take the harder second module with no ceiling. Below it you take the easier module and the band is capped at 4.0, which is why the same raw total can produce different bands.There are no high-value items. Every question in the section is worth one raw point.There is no separate timing penalty. Unanswered items simply score zero like any other missed item.The second module does not outweigh the first. The first module is the router, which makes it the more consequential of the two.
Your practice raw scores over three weeks are 27, 29, 31. Your reported band is 4.0 every time. What should you conclude?
Band 4.0 spans raw 27 to 31, so real improvement inside that range is invisible in the band. At 32 raw the band moves to 4.5. This is exactly why you should track raw score out of 47 in practice, not band.The opposite is true. Four raw points of improvement is steady progress; the band scale is simply too coarse to show it yet.The band is consistent, not inconsistent. It is a banded scale, so it holds steady across a range of raw scores by design.Nothing in the raw total tells you which task type is leaking. You need a breakdown by task type before deciding where to train.
What to record after every practice Listening section
Raw score out of 47, not just the band. The raw number moves when the band does not.
First-module percentage on its own, and whether it cleared 60 percent.
Correct answers broken down by the four task types, so you can see which one is leaking.
How many items you got wrong because of attention rather than comprehension, counted separately.
Whether any item was left blank. The target for this number is always zero.
One named reason per wrong answer, in your own words, so the pattern is visible after three sessions.
Monthly & Premium
9 more sections in this lesson
You have read the opening. The rest covers the method in full, with worked examples and practice questions that explain why each wrong answer is wrong.
All 45 lessons: Reading, Listening, Writing and Speaking
Real exam audio and record-yourself speaking drills
On our practice tests, 27 raw correct out of 47 reaches band 4.0, and band 4.0 holds until 31 raw. At 32 raw you move to 4.5.
No. A wrong answer and a blank both score zero, so guessing costs nothing and can only help. Never leave a Listening item unanswered.
It is the first-module score that decides which second module you get. At or above 60 percent you take the harder module with no band ceiling. Below it you take the easier module and your Listening band is capped at 4.0.
No. Every item in the section carries one raw point, whether it is a one-line reply item or an inference question at the end of a lecture.
Bands cover ranges of raw scores. Band 4.0 spans five raw scores, so you can gain four correct answers without the band moving, then gain the fifth and jump half a band. Track raw score in practice for a truer picture.
The overall band is the average of the section bands you took, rounded to the nearest 0.5. Across four sections, a Listening band one full point below your others costs roughly a quarter of a band overall.