How Reading Is Scored: Bands, Raw Points and Adaptive Routing
TOEFL Reading is scored from raw points, converted to a band from 1.0 to 6.0. How adaptive routing sets your ceiling and which items are cheapest to win.
11 min readMonthly & Premium
Quick answer
Every Reading item is worth one raw point, including each letter-completion blank. Your raw total across both modules is converted to a scaled figure and reported as a band from 1.0 to 6.0 in half-point steps, aligned to CEFR. Module 1 accuracy also decides whether Module 2 is harder or easier, which sets your realistic ceiling.
One point per item, and why that changes your priorities
Reading puts 50 items in front of you and each one is worth one raw point. There is no bonus for a hard question and no penalty for a wrong one. This sounds obvious and it is, but almost nobody studies as if it were true.
One detail about the real exam, because most preparation sites leave it out. ETS scores 35 of those 50 items. The rest are trial questions being tested for use in future exams, and the published blueprint does not say which ones they are. You cannot spot them while you sit there, so the right response is to treat all 50 as if they count. On our practice tests all 50 genuinely do count, because we have no questions to trial.
Consider what a single letter-completion blank is worth: one point. Now consider what the important idea question at the end of a long academic passage is worth: also one point. The letter-completion blank takes eight seconds. The important idea question takes 70 seconds and requires you to hold four paragraphs in your head. Both pay the same.
That is the whole economics of this section in three sentences. Yet the standard student instinct is to give the long, satisfying, intellectually respectable question the time it "deserves" and to hurry through the mechanical ones. That instinct costs bands. If you take one thing from this lesson, take this: value your time by points per minute, not by how serious a question looks.
The second consequence is about blanks. Because nothing is deducted for a wrong answer, an unanswered item is strictly worse than a guessed one. There is no situation, none, in which leaving an item empty is correct. Students still do it, usually out of a vague sense that a wrong answer might "look worse". It does not. It looks identical to a blank on the scoresheet, except that a blank cannot be lucky.
Bands and CEFR levels
Band
CEFR level
What it means in practice for Reading
6.0
C2
You handle unfamiliar academic topics at speed, and inference questions do not slow you down. Errors at this level are almost always careless rather than comprehension failures.
5.0 to 5.5
C1
Solid on factual and vocabulary questions. The remaining losses cluster in inference and important idea, and in the last five minutes of the section.
4.0 to 4.5
B2
You understand the passages but lose points to pace and to trap options that repeat passage wording. This is the band where technique gains the most.
3.0 to 3.5
B1
Daily Life texts are largely fine. Academic passages are where the losses are, usually because unfamiliar vocabulary forces re-reading.
2.0 to 2.5
A2
Letter-completion and short campus texts are the realistic scoring ground. Build vocabulary before you build technique.
1.0 to 1.5
A1
Below the level the academic passages are written for. Reading volume in English, daily, matters more than any test strategy at this point.
How the band scale maps onto CEFR. This is the mapping our score reports use and it is the one to plan against when a university asks for a CEFR level rather than a band.
From your answers to your band, step by step
1
Every item is marked right or wrong
Multiple choice items are marked against the single correct option. Letter-completion blanks are marked against the exact ending, so spelling counts. "thou" plus "sands" scores. "thou" plus "sand" does not, and neither does "thou" plus "sannds". This is the one place in Reading where your spelling is directly assessed.
2
Raw points are added across both modules
Module 1 and Module 2 points go into the same total. There is no separate Module 1 score on your report, even though Module 1 did something extra by deciding your route.
3
The raw total is converted to a scaled figure
The conversion is not one to one and it is not linear. Points near the bottom of the range move the scaled figure a lot. Points near the top move it less, because the top of the scale has to separate students who are all getting most items right.
4
The scaled figure is reported as a band
Bands come in half-point steps, so the scaled figure lands you in one of eleven buckets. This is why small clusters of items matter: two or three items can be the whole difference between 4.0 and 4.5, and nothing about your report will tell you it was that close.
What the adaptive routing really does to your score
Module 1 is identical for every test taker. Your accuracy on it determines whether Module 2 is the harder or the easier version. On our practice tests that decision is made at 60 percent of Module 1 correct: at or above 60 percent you route to the harder Module 2, below it you route to the easier one.
Students consistently misunderstand what this means for their score, in one of two directions.
The first misunderstanding is treating the easier route as good news. It is not. An easier Module 2 means the items you meet cannot demonstrate the upper part of the scale, because they are not built to. You can answer every question on an easy Module 2 correctly and still not land where a student who answered slightly fewer questions on a hard Module 2 lands. The route is not a difficulty adjustment made for your comfort, it is a measurement decision about which part of the scale you are being measured on.
The second misunderstanding is despair on seeing a hard Module 2. Students see harder material, conclude the test has decided they are doing badly, and their accuracy collapses for the next fifteen items. In fact a harder Module 2 is the signal that Module 1 went well. If the material in front of you gets noticeably harder after the first module, that is the best news you will get during the section.
The practical takeaway is unglamorous: Module 1 deserves your sharpest attention and your best pacing. It is the half of the section that both scores points and sets your ceiling. Warming up on Module 1 and planning to "really focus" in Module 2 is close to the worst possible strategy for this section, and it is exactly what a student who has only ever practised untimed will do.
Where your recoverable points actually are
When we look at a Reading result, the losses almost never spread evenly across the 50 items. They cluster, and the cluster tells you what to fix.
Losses concentrated in the last ten items. This is a pacing problem, not a reading problem. You know the material and ran out of clock. It is the cheapest thing to fix in this whole course and the pacing lesson is where to go.
Losses concentrated in one Complete the Words paragraph. Either the topic vocabulary was outside your range, or you were rushing at the end. Check which by looking at whether the blanks you missed were hard words or easy words. Missing "salin____" is a vocabulary gap. Missing "under____" is haste.
Losses concentrated in inference and important idea questions. This is a trap-option problem. You are almost certainly picking options that are true but do not answer the question asked, or options that go one step further than the passage licenses. Two lessons in this course address exactly that.
Losses spread thinly across factual questions. This usually means you are answering from memory of the passage rather than from the passage itself. The fix is mechanical: for every factual question, put your eyes back on the line before you commit. Students resist this because it feels slow. It is faster than being wrong.
Notice that three of those four fixes are behavioural rather than linguistic. That is the honest picture of this section for most students sitting somewhere in the middle of the scale. Your English is usually good enough for the band above the one you are getting.
Score-protection habits worth building now
Never leave an item unanswered. Nothing is deducted for a wrong answer, so a blank is a guaranteed zero you chose.
Type only the ending in Complete the Words, then read the full word back in the sentence before moving on.
Treat Module 1 as the high-stakes half. It scores points and it sets your route.
Stop judging questions by how impressive they look. Every item pays one point.
When you review a practice test, group your errors by question type before you look at anything else. The shape of the cluster is the diagnosis.
Re-do a Reading test you have already sat, two weeks later, timed. If your score jumps mostly because you remember answers, you learned nothing. If it jumps because you finished, your problem was pace all along.
Test the scoring logic
Five questions. Each one targets a scoring belief that costs students real marks, so read the reasoning for the options you rejected as well.
Two students sit the same Reading test. Priya answers 34 items correctly. Omar answers 34 items correctly. Which statement is most defensible?
Module 1 accuracy routes each student to a harder or an easier Module 2. Two students with the same raw total can have taken different routes, and the harder route measures a higher part of the scale. Identical raw totals do not guarantee identical bands.Raw points feed the conversion, but the routing means two students may not have answered comparable material. Treating the raw total as the whole story is exactly the assumption the adaptive design breaks.Question length has no bearing on value. Every item is worth one raw point, whether it is an eight-second letter-completion blank or a 70-second important idea question.Raw totals across both modules are added into one figure and converted. There is no barrier to scoring; the point is that the same figure can sit differently depending on route.
A paragraph reads: "When the ocean retre____ at low tide, these shallow basi____ trap seawater." A student types "at" for the first blank and "ns" for the second. What happens?
The sentence needs "retreats" (the subject "the ocean" takes a third person singular verb), so the ending is "ats" and "at" is marked wrong. "basins" needs "ns", which the student typed correctly. This is the single most common way students lose letter-completion points: the vocabulary is known, the grammatical ending is not checked.Letter-completion is checked against the exact ending, not against the plausibility of the meaning. "retreat" is a real word and a real reading of the stem, and it is still not what this sentence requires.Typing the ending is what the task asks for. The second answer here, "ns", is a correctly typed ending and it scores. Only the first answer fails.This reverses the marking. "at" produces "retreat" where the sentence needs "retreats", and "ns" produces "basins" which is exactly right.
With 90 seconds left you can either (a) answer four remaining letter-completion blanks by guessing endings that fit, or (b) carefully answer one important idea question. Which maximises expected raw points?
Four items at one point each is a maximum of four points. One item is a maximum of one. Even if your guessed endings are only half right, you come out ahead. This is the points-per-minute calculation that decides the last two minutes of most Reading sections.Depth of comprehension is not a scoring category here. The important idea question pays one raw point, the same as each blank you would be abandoning.Guessed answers are counted exactly like considered ones. The marker sees a typed ending, not your confidence level, and nothing is deducted for being wrong.Certainty does not increase the value of an item. A confidently correct important idea answer and a luckily correct letter-completion blank are both worth one point.
A student reports that Module 2 felt much easier than Module 1 and that they finished it comfortably. What is the most useful thing to tell them?
An easier Module 2 follows a Module 1 below the routing threshold. The easier route is not built to demonstrate the upper part of the scale, so however comfortable it felt, the realistic ceiling was set before it began. The actionable advice is to work on Module 1 accuracy and pace.Comfort is a poor proxy for score. Finishing an easier module comfortably is consistent with a modest band, and mistaking ease for success is why this pattern goes undiagnosed for months.Difficulty differences between modules are the design working, not a fault. There is nothing to disregard.Module 2 is not uniformly easier or harder. Which version you meet depends on Module 1, so the observation carries a great deal of information about how Module 1 went.
Which review habit is most likely to move a student from band 4.0 to band 4.5?
Half a band is a small number of items, and those items are almost never randomly distributed. Sorting errors by type turns a vague sense of "I am not good at reading" into a specific, fixable target, which is the only kind of target that moves a score in a few weeks.Vocabulary volume helps eventually and helps letter-completion in particular, but fifty words a week rarely intersects with the specific words on your test. It is a slow lever pulled by students who want to feel productive.Reading twice before looking at the questions doubles the most expensive part of the section and is a leading cause of unfinished sections. It converts a comprehension advantage into a timing deficit.Practising what you already do well produces confidence and no points. The items you are losing are by definition not in that set.
A word on our practice reports
When you sit a Reading test here you get a band, a breakdown by question type, and the correct answer with an explanation for every item you missed. Reading and Listening are objective, so those results are computed directly from your answers. Writing and Speaking go through expert evaluation, which is why they take longer to come back.
Use the question type breakdown rather than the headline band when you are deciding what to study next. The band tells you where you are. The breakdown tells you what to do on Tuesday.
Monthly & Premium
8 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
One raw point per item, including each individual blank in a Complete the Words paragraph. A paragraph with ten blanks is therefore worth ten raw points.
No. Nothing is deducted for an incorrect answer, so guessing is always better than leaving an item blank.
It depends entirely on the institution and the programme. Check the requirement on the university page rather than working from a general rule, and note whether they ask for an overall band, a per-section minimum, or a CEFR level.
Yes. The ending you type is checked exactly, so a misspelled or wrongly inflected ending scores zero even when you clearly knew the word.
The opposite. A harder Module 2 follows strong Module 1 accuracy and carries the higher band ceiling. If the material gets harder after the first module, that is a good sign.
Band 6.0 corresponds to C2, 5.0 and 5.5 to C1, 4.0 and 4.5 to B2, 3.0 and 3.5 to B1, 2.0 and 2.5 to A2, and 1.0 and 1.5 to A1.