The four TOEFL interview questions climb in a fixed shape: personal recall, preference, agree or disagree, then policy. Know the shape, know what is coming.
12 min readMonthly & Premium
Quick answer
Take an Interview is items 8 to 11 of TOEFL Speaking. A recorded interviewer asks four questions on one everyday topic: a personal recall question, a preference question, an agree or disagree question, and a policy or prediction question. Each answer is recorded for 44 seconds with no preparation time.
Four questions, one topic, one shape
After the seven repetition items, the screen changes. A recorded interviewer appears, introduces herself, and explains that you have agreed to take part in a short research study. Then she asks four questions about one everyday subject. In our practice sets that subject might be teamwork, outdoor activities, learning styles, leisure time, shopping habits, food, technology use, study habits, health, or travel and commuting. You answer each question into a 44 second recording that starts the instant her voice stops.
The four questions are not four independent prompts. They form a deliberate progression, from the most concrete to the most abstract, and the progression is the same every time. Knowing it is worth several points, because the hardest part of this task is not having ideas, it is organising them from a standing start. If you already know what kind of thinking each slot is asking for, you spend the first two seconds retrieving content rather than working out what is being asked.
This lesson maps the four slots: what each one is testing, what a good answer looks like in that slot, and where candidates typically leak points. The next lesson gives you a single structure that fits all four inside 44 seconds.
What the opening of the block sounds like
The first question contains two parts. What does that tell you about how to structure the answer?
The question asks what the project was and what your role was. An answer that describes the project beautifully but never says what you personally did has left half the prompt unanswered, and development is scored on whether you actually addressed the question.Both parts are part of the prompt. Dropping the first one leaves your role floating without context, which makes the answer harder to follow, not shorter.Choosing one part is a strategy for a question you did not understand. Here you understood it, and you have 44 seconds, which is enough for both parts with room for detail.A general description of teamwork is the classic underdeveloped answer. It could have been given to any question in the set, which is exactly why it sits in the middle of the scale.
When does the recording start?
There is no gap between the question and the recording. This is why the first sentence of every answer has to be something you can produce without thinking, rather than something you compose.No countdown appears. The absence of any preparation signal is the defining feature of this task and the main way it differs from older TOEFL speaking formats.You never control the recording. The test starts it, runs it for 44 seconds and ends it, which is also why you cannot finish early to save time.The interviewer video continues while you speak. It is a framing device, not a cue, and waiting for it to do something costs you seconds.
Show transcript
Hello, and thank you for taking part in this study. My name is Doctor Chen, and I am looking at how people work together in groups. I am going to ask you four questions about your own experience. Just answer as naturally as you can. Here is the first question. Think about a time you worked on a group project at school or at work. What was the project about, and what was your role in the group?
The four slots
Item
Type
Typical wording
What it tests
Most common failure
8
Personal recall
Think about a time when you... What happened, and what was your role?
Whether you can retrieve a real memory fast and tell it with concrete detail.
Telling it in generalities. "I did a group project once and it went well" uses six seconds and says nothing.
9
Preference
Do you prefer X or Y? Why?
Whether you can commit to a choice and justify it from experience.
Refusing to choose. "It depends on the situation" costs five seconds and signals no position.
10
Agree or disagree
Some people believe that... Do you agree? Why or why not?
Whether you can take a position on a general claim and support it with a reason.
Restating the claim back instead of responding to it.
11
Policy or prediction
Do you think X helps, or does it cause Y? Explain your reasoning.
Whether you can reason about effects and trade-offs under pressure, at the point where you are most tired.
Freezing, or producing thirty seconds of hedging with no claim in it.
What each question in the progression asks for, and where the points leak.
Item 8: personal recall, the easiest one to waste
This is the friendliest question in the block and it is the one most often thrown away. The prompt asks for something you actually did: a group project, a trip, a meal, a decision. The information is already in your head, complete with names and places, which means you are not being asked to invent anything. All you have to do is choose one instance and describe it specifically.
Candidates lose points here by abstracting. Asked about a group project, they say that group projects are common in university, that they have done several, that cooperation is important. Every sentence is true and none of them is an answer. The rubric wants specific personal details, and the fastest route to them is to pick a single occasion within the first three seconds and refuse to leave it.
The other trap is spending the whole 44 seconds on setup. A long description of the course, the semester and the professor leaves ten seconds for what you actually did, which is the half of the prompt that matters most. Get to your own actions by about the fifteen second mark.
One practical rule: name something. A subject, a city, a month, a number of people, a specific problem that came up. Named detail is the single clearest evidence a rater has that you are describing a real memory rather than assembling a plausible answer, and it costs no extra time to include.
Record one: the personal recall slot
Think about a time you worked on a group project at school or at work. What was the project about, and what was your role in the group?
Start speaking within one second. Your first five seconds are a short natural opener plus the choice of instance, and nothing else: something like "Yeah, the one that comes to mind is a project we did last year". While you say that sentence, your brain is picking which project. Do not wait until you have chosen to begin talking. By ten seconds you should be saying what the project was, and by twenty you should be on what you personally did.
0:44
Model answer
The opener buys time without being empty. "Yeah, the one that comes to mind is" takes about two seconds, sounds like a person thinking, and commits to a single instance. It is not a memorised flourish and it does not restate the question.
The instance is named immediately. A marketing course, last year, four people, a local bakery. Four concrete details inside the first twelve seconds, which is what the top of the scale means by specific personal detail.
The role is answered directly. "My role was mostly the research side" lands around the twenty second mark and then gets its own supporting detail. The prompt asked two things and the answer visibly does both.
There is a small narrative turn. The finding about the morning trade gives the answer somewhere to go, which is what separates a developed answer from a longer description. It also produces a natural ending rather than a stop.
The discourse markers are real ones. "Yeah", "so", "honestly" are what people actually say while organising a thought out loud. They fill the joins with sound rather than silence, which is what keeps a thinking pause from reading as a breakdown.
Item 9: preference, where hedging kills you
The second question puts two options in front of you and asks which you prefer. Do you like taking the lead or following directions. Do you prefer studying alone or in a group. Do you prefer cooking at home or eating out. The structure is always the same and so is the trap.
The trap is "it depends". It is an honest answer, it is often the truest answer, and in a 44 second window it is a disaster. It costs you three to five seconds, it gives the rater nothing to follow, and it usually leads into a balanced discussion that never lands anywhere. The scale rewards a developed position. It does not reward accuracy of self-description.
So commit in the first sentence. You are allowed to add nuance later, and adding it later is actually the strongest move available: state the preference plainly, give the reason, give the instance, and then in the last ten seconds add the condition under which you would choose the other one. That gives you a clear position and visible nuance, in the order that makes both of them count.
The second failure here is answering with a principle instead of a preference. "I think good leadership is important in any group" is not a statement about what you prefer. Use "I", use a verb of preference, and make it unambiguous within the first eight seconds.
Item 10: agree or disagree, where you need a reason not a restatement
The third question presents a general claim, usually attributed to "some people", and asks whether you agree. Some people believe group projects build skills that matter professionally. Some people think technology has made studying easier. Some people say home cooked food is always healthier.
This is the first genuinely abstract slot, and the most common weak answer is the mirror answer: repeating the claim in slightly different words and calling that agreement. "Yes, I agree that group projects help students develop skills, because they develop skills that help students" is a real pattern and it is completely empty. The rater is listening for a reason that adds information the prompt did not contain.
The reliable move is to agree or disagree in one short sentence and then immediately go to the mechanism. Not "I agree because it is useful", but "I agree, and the reason is that you end up having to explain your work to people who do not already understand it, which is most of what a job actually is." The mechanism is the part that cannot be recovered from the prompt, so it is the part that demonstrates thinking.
You may agree partially, but say so decisively. "I mostly agree, with one exception" is a strong, clear position. "Well, there are arguments on both sides" is not, because it delays commitment past the point where the rater needed to hear one.
Item 11 in the interviewer voice
What makes this the hardest slot in the block?
The fourth slot asks you to weigh two possible effects against each other, which is abstract work, and it lands after three answers have already used your concentration. That combination, not the language, is what makes it the low scoring item for most candidates.The words are ordinary: teams, communication, progress. Nothing in the phrasing is above the level of the earlier questions.Every one of the four answers is recorded for 44 seconds. The time limit is identical across the block.No outside knowledge is expected or rewarded. Your own experience of working with other people is entirely sufficient material.
The question offers two alternatives. What is the strongest opening move?
A clear claim early gives the rest of the answer something to develop, and development is what the upper part of the scale is measuring. Nuance can be added in the final ten seconds without weakening the position.A balanced view with no choice reads as undeveloped, because there is no claim for the supporting detail to attach to.Repeating the question back burns seconds and adds no information the rater did not already have. It is one of the most reliable ways to end up with a short, thin answer.Saying both are true and stopping is the hedge in its purest form. It is defensible as a view and it produces exactly the vague, underdeveloped recording the middle of the scale describes.
Show transcript
Last question. Do you think working in teams helps people improve their communication skills, or do you feel it slows down individual progress? Explain your reasoning.
Record two: the policy and prediction slot
Do you think working in teams helps people improve their communication skills, or do you feel it slows down individual progress? Explain your reasoning.
This is the slot where people freeze, so your first five seconds have to be mechanical. Start with a two second acknowledgement of the trade-off, then pick a side inside the same breath: "I can see both, but I would say it mostly helps." That sentence is producible before you have thought of any reason at all, and saying it out loud usually generates the reason. Then go straight to a mechanism, and use an example from your own experience rather than a general claim.
0:44
Model answer
The position arrives in the first four seconds. "I can see both sides, but I would say it mostly helps" acknowledges the trade-off and commits, in one sentence. That is the whole hedging problem solved without losing the nuance.
The reason is a mechanism, not a label. Instead of "because communication is important", the answer explains how the effect happens: alone you are never corrected, in a team you are corrected quickly. That is information the prompt did not contain, which is what the reasoning part of the question is asking for.
A personal instance grounds it. The research notes example takes about twelve seconds and converts an abstract claim into a specific one. This is the same move that works in item 8, reused in the hardest slot.
The concession comes last and strengthens the answer. Conceding the short term cost after the position has been established shows range without ever putting the claim in doubt. Doing it first would have read as indecision.
The close is real speech. "But the skill you pick up is the one that keeps mattering later" is a natural ending, not a formula. Compare it with a memorised closer like "in conclusion, for these reasons I believe", which spends three seconds restating what you already said instead of adding the last piece of content.
A slot by slot game plan
1
Item 8: pick one instance in three seconds and name it
Open with a natural lead in while you choose. Name a course, a place, a month or a number of people inside the first twelve seconds. Reach your own actions by about twenty seconds. Do not describe the general category of thing, describe the one that happened.
2
Item 9: commit in the first sentence, qualify at the end
Say which one you prefer using "I" and a preference verb. Reason, then instance. Save any nuance for the last ten seconds, where it reads as range rather than as indecision.
3
Item 10: agree or disagree, then explain the mechanism
One short sentence of position. Then the how, not the what. If your reason could be produced by someone who had only heard the question and knew nothing else, it is not yet a reason.
4
Item 11: acknowledge, commit, mechanise, concede
Two seconds acknowledging the trade-off, a claim in the same breath, a mechanism, one personal instance, and a short concession at the end. This sequence works even when you have no opinion at all, which is the point of having it ready.
5
Between items: reset in two seconds
The next question starts whether you are ready or not. Whatever happened in the last answer is finished and separately scored. Unclench, breathe out, and listen. Carrying item 10 into item 11 is how one weak answer becomes two.
Before you call an interview answer finished
I answered every part of the question, including the second half of two part prompts.
My position or instance was clear within the first ten to twelve seconds.
There is at least one named, concrete detail that only I could have supplied.
My reason explains a mechanism rather than repeating the prompt in new words.
I used between about 40 and 43 of the 44 seconds.
The answer ends on a complete thought rather than being cut off mid clause.
Nothing in the answer sounds like a memorised phrase inserted to seem academic.
How to practise the block
Practise the four answers as a block, not as individual questions. The difficulty of item 11 is partly that it comes fourth, and you cannot rehearse that by doing item 11 first while fresh. Set up four prompts, hit record, and go through all four with no stopping and no notes, exactly as the test runs.
Then listen back to all four in a row and look for the shape of your own decline. Most candidates find that answer one is a little too long on setup, answer two hedges, answer three restates, and answer four is thirty seconds instead of forty three. Those four faults are each fixable and none of them is a language problem, which is why this block responds so quickly to structural practice.
One last piece of realism. Everybody has a bad fourth answer sometimes, including strong candidates. It costs you one item out of eleven. The candidates it really damages are the ones who let it become a story about how the test went, and then carry that into their next practice session. Score it, note the fault, and run the block again tomorrow.
Monthly & Premium
12 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
Four questions, items 8 to 11 of the Speaking section. Each answer is recorded for 44 seconds, and the recording starts automatically when the question audio ends. There is no preparation time and no way to pause, restart or finish early.
The progression is consistent: a personal recall question, a preference question, an agree or disagree question, then a broader policy or prediction question. The topic changes between test sets but the shape of the climb does not, which is what makes it worth preparing.
No. The interviewer is a recording and cannot hear you. If you miss part of a question, answer the part you caught, confidently and on topic. A clear answer to most of the question scores far better than a paralysed silence while you try to reconstruct all of it.
Nobody is checking. What matters is that the answer sounds like a real memory, which in practice means specific, and true memories are simply the fastest source of specific detail under pressure. Inventing a story costs you thinking time you do not have.
Say the acknowledgement and pick a side anyway: "I can see both, but I would say it mostly helps." You are being scored on how you develop a position, not on whether it is your considered lifelong view. Saying the sentence out loud very often produces the reason that follows it.
Vary them, though not because a repeated opener is penalised on its own. The reason is practical: one opener will not fit all four question shapes, and forcing it onto a question it does not suit costs you seconds and pushes you toward an answer you did not mean to give. Have two or three natural lead ins per slot and rotate, so each answer starts quickly and starts on the actual question.