The first 18 minutes of Reading decide what the rest of it looks like

What actually happens

You do not sit one continuous Reading section. You sit a router — a first block of questions everybody gets — and then, depending on how the router went, the test gives you either a lower module or an upper module for the rest.

Writing and Speaking are not adaptive — everybody on a given form gets the same tasks. This is also why total test time is published as a range, roughly 1 hr 23 min to 1 hr 29 min, rather than a fixed number. Which module you get changes how long you sit there.

Why this matters more than it sounds

The intuitive read is “so the test gets harder if I do well” — true, and not the important part. The important part is that the two stages are not equivalent, and the router is the one that decides your range. A candidate routed to the lower module is answering questions drawn from a different difficulty band. Doing brilliantly there does not put the upper module's score range back within reach.

Two things you cannot know, so stop trying

You cannot tell which module you got. There is no announcement, no label, no visible difficulty marker. People will tell you they could feel it; they are guessing, and the guess is expensive — a candidate who decides mid-section that they have been routed low tends to either coast or panic, and both cost real points on questions that are still being scored.

Related, and it catches people the same way: not every item you answer is scored. Reading is 50 items but 35 raw points; Listening is 47 items but 35. ETS uses the remainder to gather performance data on questions being trialled for future tests. You cannot identify them — they are not marked, not grouped, and not distinguishable by type or difficulty. Any minute spent deciding a question “does not count” is a minute spent on something unknowable, and the cost of guessing wrong is that you skipped a scored item.

Why reading this does not fix it

Pacing is the one thing on this page you cannot practise by understanding it.

You now know the router carries disproportionate weight and that you should arrive warm. That knowledge changes nothing about what your hands do at minute four, because the failure is physical — you settle, you re-read a line, you check a question twice, and eleven minutes in you are behind without having noticed.

Untimed practice actively hides this. Work through a passage set with no clock and you will finish it accurately, feel fine, and learn nothing about the only skill the two-stage format really tests. The clock is not a difficulty setting you add at the end. It is the thing being measured.