TOEFL Speaking: Mastering the Four Tasks and Their Tiny Clocks

TOEFL Speaking compresses four tasks into about 16 minutes, with preparation windows measured in seconds and responses capped around a minute. Nobody composes good answers in 20 seconds; successful candidates walk in with the answer shapes already built and spend their seconds on content. This guide builds those shapes.

Where Speaking's 0 to 30 comes from, and what raters listen for, is covered in the TOEFL scoring guide. If IELTS's face-to-face interview suits you better, that trade-off is exactly the one our exam comparison weighs.

Task 1: the independent opinion

One question about your preference or opinion ("Do you agree that children should learn a musical instrument?"), roughly 15 seconds to prepare, 45 to speak. The shape: position in the first sentence, one reason developed with a specific example, a second reason if time allows. Pick the side you can support fastest; raters score development and delivery, not sincerity.

The classic failure is the list: three reasons, none developed, each one sentence long. One reason with a concrete example ("when my brother learned piano, his focus in school improved within a year") outscores three assertions every time.

Tasks 2 and 3: read, listen, connect

Both integrated tasks follow the same skeleton: a short reading (45 to 50 seconds on screen), then a listening that responds to it, then a question that asks you to connect the two. Task 2 is campus life: an announcement plus a student's reaction. Task 3 is academic: a concept definition plus a lecture example.

The scoring insight is that the listening carries the answer. The reading sets context; the speaker's opinion or the professor's example is what your response must reproduce. Notes should be split accordingly: a few words for the reading's claim, most of your attention on the listening's reasons.

Reusable shapes:

• Task 2: "The announcement says X for reasons A and B. The student disagrees. First, she argues... Second, she points out..." Naming the speaker's two reasons, with their supporting details, is essentially the whole rubric.
• Task 3: "The term is X, which means... The professor illustrates this with... " Then walk the example through, connecting it back to the definition at the end.

You are not asked for your opinion in either task; giving one spends seconds the rubric does not pay for.

Task 4: the lecture summary

Listen to a short academic lecture, then summarize it with about 20 seconds of preparation and 60 to speak. The lecture is always structured, typically one concept with two examples or two contrasting approaches, and your response should mirror that structure: concept first, then each example with its key detail. The note-taking system that serves the Listening section is the same one that wins Task 4: capture the skeleton, not sentences.

What the raters actually score

Responses are evaluated on three dimensions: delivery (clear, sustained speech at a natural pace), language use (grammar and vocabulary range deployed accurately) and topic development (complete, coherent content within the time). ETS uses certified human raters together with automated scoring.

Practical consequences: a small stumble is fine and recovery matters more than perfection; pace is worth rehearsing because rushed speech damages delivery scores more than a modest accent ever will; and finishing your structure matters, so if the clock is closing in, compress the final point rather than abandoning it mid-sentence.

The practice loop

Speaking improves on the same loop as writing: attempt under real timing, get criterion-level feedback, fix one thing, repeat. Record every response; the played-back version contains information the speaking version hides. Run full four-task sets at least weekly once your TOEFL plan enters its timed phase, so the 16-minute rhythm itself becomes familiar.

For the feedback step between tutor sessions, Verbola's speaking practice transcribes and scores responses on delivery, language use and topic development; as with all AI feedback, including ours, treat the score as a practice estimate pointing at your weakest dimension, never as an official result. Four weeks of shaped, reviewed repetitions is typically enough to turn the section's tiny clocks from a threat into the reason the section feels easy: you only ever have to fill 60 seconds, and you arrive knowing exactly how.