TOEFL Speaking Mastery Guide
All 4 tasks explained with templates, the full scoring rubric decoded, AI scoring demystified, and the 6 common mistakes that cost test-takers points.
Last updated: 2026 · 16 min read
Section Overview
The TOEFL Speaking section has 4 tasks completed in 16 minutes total. Task 1 is Independent — all ideas come from you. Tasks 2, 3, and 4 are Integrated — you read and/or listen to source material, then summarize or synthesize it.
| Task | Type | Prep Time | Speak Time | Input |
|---|---|---|---|---|
| Task 1 | Independent | 15 seconds | 45 seconds | Personal topic — no reading or audio |
| Task 2 | Integrated | 30 seconds | 60 seconds | Campus announcement (read) + conversation (listen) |
| Task 3 | Integrated | 30 seconds | 60 seconds | Academic concept (read) + professor lecture (listen) |
| Task 4 | Integrated | 20 seconds | 60 seconds | Professor lecture (listen) — no reading |
Responses are recorded through the computer's microphone and sent to ETS for scoring. You will hear a beep when your preparation time ends and speaking begins. Another beep signals the end of your speaking time. Speaking after the final beep is cut off and not scored.
Scoring Rubric
Each Speaking task is scored on three criteria, each on a 0–4 scale. The raw score is combined and scaled to the final 0–30 Speaking section score.
Clear speech, natural pacing, minimal hesitation, intelligible throughout.
Mostly clear. Some minor hesitation or pronunciation issues, but does not impede understanding.
Noticeable hesitation, unnatural pacing, or pronunciation problems that sometimes impede understanding.
Frequent hesitation, unclear pronunciation, very difficult to understand.
Varied sentence structures, accurate grammar, precise vocabulary with minimal errors.
Mostly accurate. Some grammatical errors but they do not impede meaning. Adequate vocabulary range.
Limited vocabulary, repetitive structures, or errors that sometimes obscure meaning.
Frequent errors that significantly impede communication.
Clear progression, ideas fully developed, well-organized with logical connections between ideas.
Mostly coherent. Some ideas not fully developed but overall task is addressed.
Ideas incomplete, poorly organized, or the task is only partially addressed.
Ideas very limited, incoherent, or the task is largely not addressed.
How AI Scores Your Speaking Response
As of recent updates, ETS uses SpeechRater — an automated speech scoring engine — as the primary or sole scorer for TOEFL Speaking, with human raters used for quality control and edge cases. This is a significant change from the traditional dual-human rater system.
What SpeechRater measures
- Speech rate: Words per minute, with pauses excluded. Too slow or too fast both lower Delivery scores.
- Pause frequency and duration: Long pauses mid-sentence signal disfluency. Pauses between sentences are acceptable.
- Pronunciation accuracy: Phoneme-level scoring compared against a model. Intelligibility matters more than accent.
- Repetition and filler words: Frequent "um," "uh," "like," "you know" lower fluency scores.
- Grammar error detection: Identified through language model analysis of the transcript.
- Lexical diversity: Variety of vocabulary used. Repeating the same 5–10 words throughout lowers Language Use scores.
- Content relevance: Whether the response addresses the prompt — measured by keyword and semantic matching.
Implications for preparation
- Do not try to memorize and recite a pre-written response. AI detects unnatural fluency patterns and repetitive language from over-rehearsed scripts.
- Speak at a moderate, consistent pace (120–160 words per minute for most speakers). Do not rush to pack in more content.
- Reduce filler words actively. Practice pausing silently instead of saying "um" — a brief pause scores better than a filler word.
- Use varied vocabulary. Vary your sentence structures. Avoid starting every sentence with "I think" or "It is."
- Address the task directly and completely. Off-topic responses score poorly even if they are linguistically excellent.
Task 1 — Independent (15s prep / 45s speak)
Task 1 asks for your personal opinion, preference, or experience on a familiar everyday topic. There is no reading passage or audio — all ideas come from you. Common prompt types:
- Preference: "Some people prefer X; others prefer Y. Which do you prefer and why?"
- Agree/Disagree: "Do you agree or disagree with the following statement? [Statement]"
- Description: "Describe a person/place/experience that [quality]. Explain why."
Strategy
- Choose ONE clear position in your 15-second prep time. Do not try to present both sides.
- Outline two supporting points. Keep each point simple enough to explain in 15–20 seconds.
- Use a specific example or personal experience to support at least one point. Generic statements ("it is very helpful") score lower than specific ones ("when I had to prepare a presentation at work, I found that...").
- Do not change your position mid-response — this signals poor organization and confuses the scorer.
- Aim to speak for the full 45 seconds. If you finish early, you have not developed enough content.
Task 2 — Campus Announcement (30s prep / 60s speak)
You read a campus announcement (75–100 words, 45 seconds to read) about a policy change or proposal. You then listen to a conversation (60–80 seconds) in which a student reacts to the announcement. In your response, you must:
- Briefly describe what the announcement says
- State the student's opinion (positive or negative)
- Explain the student's two main reasons for that opinion
During preparation
While reading the announcement, note: (1) the change or proposal, (2) the stated reason given by the institution. While listening to the conversation, note: (3) the student's overall reaction (favor/oppose), (4) their first reason, (5) their second reason, with any specific details.
Do not state your own opinion. The task asks for the student's view, not yours. This is the most common mistake on Task 2.
Task 3 — Academic Concept (30s prep / 60s speak)
You read a short academic passage (75–100 words, 45 seconds) that defines or describes a concept from a university course. You then listen to a professor give a lecture example that illustrates the concept. In your response, you must:
- Define the academic concept from the reading
- Explain the professor's example(s) from the lecture
- Show explicitly how the example illustrates the concept
Strategy
From the reading: underline or note the concept name and its key defining feature(s). From the lecture: note the example (who/what/where) and how it demonstrates each defining feature you noted. In your response, connect each example detail back to the definition. The explicit connection ("this shows the concept of X because...") is what earns the highest Topic Development score.
Task 4 — Lecture Summary (20s prep / 60s speak)
You listen to a 90-second academic lecture on a topic from any field. There is no reading passage. The lecture typically presents a main concept and two sub-points, examples, or aspects. In your response, you must:
- Introduce the overall concept the professor discussed
- Explain the first sub-point or example with relevant details
- Explain the second sub-point or example with relevant details
Strategy
Professors in Task 4 almost always present exactly two examples or two supporting points. Structure your notes during listening: main concept at top, then Point 1 with its details, then Point 2 with its details. Use your 20-second prep to confirm your notes cover both points.
Your 60-second response maps directly onto your notes: "The professor discussed [concept]. One point was... For example... Another point was... As the professor explained..."
Response Templates
Templates are structural frameworks — not scripts. Use them to organize your thoughts during prep time, then speak naturally. Over-reliance on memorized language patterns is detected by AI scoring and penalized.
"I prefer [position] for two main reasons. First, [Reason 1 + brief example]. Second, [Reason 2 + brief example]. For these reasons, I believe [restate position briefly]."
Fill in the brackets with your specific content. The restatement at the end is optional if you are running low on time.
"The announcement states that [summary of change/proposal]. The student [agrees/disagrees] with this because [Reason 1] and [Reason 2]. Specifically, the student points out that [detail from conversation]."
Keep the announcement summary to 1–2 sentences. Spend the majority of your 60 seconds on the student's reasons.
"The reading describes [concept name] as [brief definition]. The professor illustrates this with [example]. In the example, [explain what happens] — this shows [connection to definition] because [why it matches the concept]."
The explicit connection phrase ('this shows X because...') is what differentiates a 3-score from a 4-score response.
"In the lecture, the professor discusses [main concept]. The first point is [Point 1] — for example, [detail]. The second point is [Point 2] — the professor explains that [detail]. Together, these show [brief wrap-up]."
The wrap-up is optional. Use it only if you have time remaining. Never rush Points 1 or 2 to fit in a wrap-up.
Common Mistakes That Cost Points
ETS SpeechRater detects unnatural fluency patterns consistent with recitation. If your response sounds like you are reading from memory, your Delivery score drops. Prepare structures, not scripts.
If you are still speaking when the final beep sounds, your sentence is cut off mid-thought. This damages Topic Development scores significantly. Practice stopping at exactly 45 or 60 seconds.
Spending 5–8 seconds saying 'Well, um, I think that...' before your main point wastes 10–17% of your speaking time. Start with your position or the first fact immediately.
'Um,' 'uh,' 'like,' 'you know,' 'basically' are all tracked by SpeechRater and reduce your fluency score. Replace them with a brief silence. A clean pause scores better than a filler word.
Tasks 2, 3, and 4 ask you to summarize source material — not share your personal view. Adding your opinion wastes speaking time and suggests you misunderstood the task.
Saying 'I prefer A because it is better and more convenient' without any specific example or elaboration will earn at most a score of 2 on Topic Development. Always give one concrete supporting example.
Practice Tips
- Record every practice response on your phone or computer. Listen back immediately and identify: filler words, pacing issues, incomplete ideas, and any sentences that were grammatically wrong.
- Time yourself strictly on every practice — use a stopwatch. Speaking for exactly 45 or 60 seconds (not 40, not 65) should become automatic.
- Practice Task 1 with random prompts daily. Use the question types above as a guide, or find TOEFL practice prompt lists online. The goal is to be able to generate two developed supporting points within 15 seconds for any topic.
- For Tasks 2–4, practice note-taking speed. Your notes from prep time are your script — the more organized and complete they are, the better your response will be.
- Shadow native speakers to improve pronunciation and natural rhythm. Listen to a 20-second clip, then immediately repeat it at the same speed and with the same rhythm. This trains your mouth to produce natural English prosody.
- Get AI feedback. Tools like FullPracticeTests score your Speaking responses on the full rubric and identify exactly which criterion is pulling your score down.
Ready to practice TOEFL Speaking with AI feedback?
Take a Free Practice Exam →No sign-up required · Full exam · AI scoring available