Sin taxes work. Mexico's 2014 soda tax cut consumption by about 8 percent in the first year and even more in low-income households. Yes, the burden is regressive, but the health gains for those same households are larger because they had higher baseline consumption. The tax revenue can also fund clinics in the neighborhoods that pay it.
Writing
Academic Discussion
2026 Task Type
TOEFL Academic Discussion 2026: Band 6 Framework, Rubric, and 3 Annotated Sample Responses
The Academic Discussion task is the longer of the two writing tasks introduced in the January 2026 TOEFL iBT redesign. You read a professor's question and two student posts, then add your own contribution of roughly 100 words in 10 minutes. It looks deceptively simple — it is structurally a forum reply, not an essay — but the band ceiling on this task is decided in the first 30 seconds, when you choose how to position your stance against the two student posts already on the screen. This guide breaks down what raters reward, the four-part framework that lands at band 6, the 1-6 rubric in plain language, and three fully worked sample responses so you can see exactly what each band looks like in practice.
1. What the Academic Discussion task tests
The screen shows a course discussion board. A professor poses a question — usually a policy question, a technology question, or a values question with a defensible position on either side. Two students have already posted brief responses, typically 50 to 60 words each, taking different angles. Your job is to add the next post: take a clear position, support it with one specific reason or example, and engage with at least one of the previous posts.
ETS is testing whether you can participate in an academic conversation in real time. That means raters care about three things in roughly equal measure: do you take a clear position, do you support it with reasoning that is specific rather than generic, and do you engage with the existing thread instead of writing in isolation. A response that ignores Paul and Diana and writes a standalone mini-essay will cap at band 4, no matter how clean the grammar is.
A note on what graders actually do: Each Academic Discussion response is evaluated against the 1-6 rubric by trained raters working from the official ETS scoring guide. They do not look for a particular structure or buzzword; they look for whether your post would advance the conversation if it were a real classroom thread. That mental model — "would the professor read this and find it useful?" — is the single most reliable check before you submit.
2. How it differs from the old Independent Essay
If you are coming from older TOEFL prep materials, you may be drilling the five-paragraph Independent Essay format. Drop it. The Academic Discussion task is the spiritual successor to that essay, but the structural and length differences are large enough that an essay-style response is actively penalized.
| Dimension | Old Independent Essay (pre-2026) | Academic Discussion (2026) |
|---|---|---|
| Time | 30 minutes | 10 minutes |
| Word target | 300+ words | 90 to 130 words |
| Structure | Intro, 3 body paragraphs, conclusion | Single forum post, 1-2 paragraphs |
| Engagement | None — standalone argument | Required — must reference at least one prior post |
| Voice | Formal academic essay | Conversational but academic — like a smart Reddit comment |
The single biggest mistake students make is bringing essay habits to this task: thesis statements, "in this essay I will argue," topic sentences, conclusion paragraphs. None of those fit a 100-word forum post. Strip them out and you free up 40 percent of your word count for actual reasoning.
3. The 1-6 rubric, decoded
The Academic Discussion task is scored on the same 1.0 to 6.0 band scale as the rest of the Writing section, in 0.5 increments. Raters score four dimensions: position clarity, support quality, engagement with the thread, and language control. The first three are content; the fourth is grammar, vocabulary, and flow. Below is what each band actually looks like in practice — written in the language of decisions, not in ETS-speak.
| Band | What it looks like |
|---|---|
| 6.0 | Clear position in sentence one. One specific example or piece of reasoning, not a generality. Names at least one of the previous posts and engages with its argument. Acknowledges the strongest counter and resolves it. Varied sentence structure, no errors that obscure meaning. |
| 5.0 | Clear position. Solid example. References a previous post. Misses the counter-argument or handles it superficially. Minor language slips. |
| 4.0 | Position is there but takes two sentences to find. Example is generic ("studies show", "experts agree"). Acknowledges the thread but does not engage with specifics. Some grammar drift but meaning is clear. |
| 3.0 | Position is unclear or shifts mid-post. No specific support. Treats the task like a standalone essay; ignores the other students. Repeated grammar errors visible. |
| 1-2 | No clear position, or the response is off-topic, or under 50 words. Errors that make the message hard to follow. |
The dirty secret of this rubric: most students who plateau at band 4 do so because their support sentence is generic. "Studies have shown that exercise is good for mental health" earns no content credit because there is no example, no causal mechanism, and no engagement with the thread. Replacing one generic sentence with one specific instance — a class you took, a friend's experience, a news event — is often the difference between band 4 and band 5.
4. The CLAIM-SUPPORT-COUNTER-RESOLVE framework
Every band 6 response we have analyzed fits inside the same four-step shape. Memorize the moves, not the words. On test day you read the prompt and slot your content into each step.
Step 1 — Claim (1 sentence)
Open with your position. No "in my opinion," no "I would like to argue." Just state what you think. Reference one of the two previous posts here if it strengthens the framing.
Step 2 — Support (2-3 sentences)
Give one specific reason or example. Specific means a place, a person, a class, a number, a moment in time — not "many people" or "research shows." This is where 60 percent of your word count lives.
Step 3 — Counter (1 sentence)
Acknowledge the strongest objection to your position. Often this is one of the previous posts. Phrase it fairly, do not strawman.
Step 4 — Resolve (1 sentence)
Explain why your position still stands despite the counter. This is the move that pushes you from band 5 to band 6 — most band 5 responses skip it entirely.
That gives you 5 to 6 sentences total, which lands naturally between 100 and 120 words. If you finish at 80 words you are missing the resolve; if you push past 140 you are padding the support.
5. The 10-minute pacing plan
You get exactly 10 minutes. The clock starts the moment the task loads, while you are still reading the prompt. This allocation works:
- 1 0:00 to 1:30 — Read. Read the professor's question carefully. Read both student posts at normal speed. Do not skim — you need to know which one to engage with.
- 2 1:30 to 2:30 — Plan. Pick your stance. Pick your specific example (have one ready before you start typing). Decide which student post you will name. Jot two words for your counter.
- 3 2:30 to 8:30 — Draft. Run the four-step framework. Type without going back to edit. If a word escapes you, write a simpler one. Aim to finish around the 8:30 mark.
- 4 8:30 to 10:00 — Proofread. Read once for content (does each step land?), once for grammar (subject-verb agreement, articles, tenses). Fix the cheapest errors first.
Common timing trap: Spending 4 minutes deciding your stance because both options seem equally defensible. They are. Pick the one with the easier specific example. Raters do not score the wisdom of your position — they score how well you defend it.
6. Sample 1: Sin taxes (band 6)
Prompt: Your professor is teaching a public policy class. Read the question and the responses of two students. Write a post that contributes to the discussion in your own words. An effective response will contain at least 100 words.
Professor Hayes: Some governments tax unhealthy products like sugary drinks, alcohol, and tobacco at higher rates than other goods. Supporters say these "sin taxes" reduce consumption of harmful products and fund public health programs. Critics say they unfairly target lower-income consumers and rarely change behavior. Should governments use sin taxes? Why or why not?
I disagree. Sin taxes are paternalistic and they do not change long-term behavior. People shift to cheaper, often equally unhealthy alternatives. The real driver of consumption is marketing and availability, not price. Regulating advertising and soda machines in schools would do more than another tax.
I side with Paul on this one, but with one caveat. Mexico's soda tax is the cleanest natural experiment we have, and the consumption drop in lower-income households was both real and durable through year three of the program. That answers Diana's "people switch" objection in this specific case — the cheaper-but-equally-bad substitute did not materialize at the volume critics predicted, partly because the tax also priced up the cheaper store-brand sodas that would have been the obvious switch. Where I would push back on Paul is the regressivity claim. The fix is not to drop the tax; it is to ring-fence the revenue for primary care in the same neighborhoods, which Mexico has been inconsistent about. Done that way, sin taxes are progressive in net effect.
Why this scores band 6: Position in sentence one ("I side with Paul"). Specific support — names Mexico, gives the year-three durability point, addresses the substitute argument with a concrete mechanism. Engages with both Paul and Diana by name and addresses their actual arguments rather than caricatures. Counter is acknowledged ("regressivity claim") and resolved (ring-fenced revenue). 127 words, varied sentence structures, no errors. Reads like a real graduate-seminar comment.
7. Sample 2: AI in classrooms (band 6)
Prompt: Some universities are integrating AI assistants into their classrooms. Supporters say these tools personalize learning. Critics say they erode core skills. Should AI assistants be a regular part of university coursework?
Professor Owen: Universities are increasingly experimenting with AI assistants in coursework. Should these tools be a regular part of how students learn at university? Read the two posts below and add your own perspective.
I think AI assistants are great for learning. They give every student a personal tutor that is available at 2am, which is when most coursework actually gets done. They can also explain a concept in five different ways until something clicks. That is something a professor with 200 students cannot offer.
The convenience is real but it comes at a cost. Students who use AI to draft early outlines never develop the muscle for organizing their own thoughts. The skill universities are supposed to build — generating ideas under pressure — atrophies. We risk producing graduates who are very good at editing AI output but cannot write a paragraph from scratch.
Both Mei and Kwame are right about different stages of learning, which is why "should AI assistants be regular" is the wrong question. The right question is when in a course they should appear. In a sophomore writing class I took last year, the professor banned AI for first drafts but encouraged it for revision. By the end of the semester, our drafting was sharper because we still trained that muscle, and our revising was faster because we used AI well. Mei's 2am tutor is real value, and Kwame's atrophy concern is real risk — they coexist depending on the assignment. Universities should design that boundary at the syllabus level, not leave it to each student.
Why this scores band 6: Reframes the question rather than picking a side, which is a sophisticated band 6 move. Names both Mei and Kwame and engages with their actual arguments. Specific example (the sophomore writing class with the drafts-vs-revision split) carries the support load. The resolve sentence ("Universities should design that boundary at the syllabus level") is concrete enough that a professor would want to discuss it. 122 words, no errors.
8. Sample 3: Remote work (band 4, annotated)
Prompt: Should companies require employees to come into the office, or allow them to choose remote work? Two students have weighed in below.
Professor Ortiz: Many companies are deciding whether to require in-office attendance or to allow remote work. What is your view?
Remote work is more productive and more humane. The commute alone returns hours per week to employees. Companies that insist on offices are mostly trying to justify their leases.
Office time matters because that is where mentorship happens. Junior employees learn by overhearing how senior colleagues handle hard conversations. You do not get that on Zoom.
I think remote work is a good idea in many cases. There are many advantages of working from home such as flexibility and saving time on commuting. Studies have shown that employees who work from home are often more productive. However, working in office is also important because employees can communicate face to face. In conclusion, both options have benefits and companies should let employees choose what is best for them.
Why this scores band 4 (and how to push it to 5): The position is technically there but takes the whole post to land ("companies should let employees choose"). The support is generic — "studies have shown" with no study, "many advantages" with no specifics. Lina and Tom are not named or engaged with. The "in conclusion" essay habit signals the writer brought the wrong format to the task. To push this to band 5: cut the "studies have shown" line, add one specific example (a job, a project, a real friction point), and reference Lina or Tom by name. Word count is fine at 89 — content density is the bottleneck.
9. Common mistakes that cost points
Treating it like an essay
"In this post, I will argue that..." is a thesis statement. Thesis statements are for essays. The Academic Discussion task is a forum reply. Open with the position itself, not a meta-sentence about what you are going to argue.
Ignoring the other students
Writing a standalone post that never references Paul, Diana, or whichever names are on screen is the single fastest way to cap at band 4. Engagement with the thread is one of four scoring dimensions. You can disagree, agree, partially agree, or reframe — but you must address them.
Generic support
"Studies have shown," "experts agree," "many people believe" — none of these earn content credit because they cannot be falsified or visualized. Replace each with a specific instance: a class, a job, a country, a year, a person.
No counter, no resolve
A response that only argues the pro side reads like a one-sided pitch. A band 6 response acknowledges the strongest objection and explains why the original position survives it. Skipping this step is the single most reliable way to cap at band 5.
Pasting from the prompt
Copying a long phrase from the professor's question or from a student post inflates word count without earning any content points. Raters are trained to spot it and many automated checks discount copied text from the word count entirely. Paraphrase briefly, then move to your own contribution.
Skipping the proofread
A clean draft becomes a band 6 with a 90-second proofread; without it, it caps at band 5. Subject-verb agreement, article use ("a" vs "the"), and tense consistency are the three error types that are cheapest to catch and cost the most when missed. Always leave the last 90 seconds for these.
10. FAQ
How long should a TOEFL Academic Discussion response be?
Aim for 90 to 130 words. Band 5 and 6 responses typically land between 100 and 120 words. Going under 80 words almost always caps you at band 3 because there is not enough content for raters to evaluate; pushing past 140 usually means you are repeating yourself or padding rather than reasoning.
How much time do I get for the TOEFL Academic Discussion task?
You get 10 minutes total. The cleanest pacing is 1 minute to read the prompt and the two student posts, 1 minute to plan your stance, 6 to 7 minutes to draft, and 1 to 2 minutes to proofread. The on-screen timer counts down from the moment the task loads, so you cannot skip planning to get more drafting time.
Do I need to disagree with the other students in my Academic Discussion response?
No. You can agree with one student, agree with both, disagree with both, or take a different angle entirely. What matters is that you take a clear position and support it with a specific reason. Raters reward responses that engage with at least one of the previous posts, but they do not reward picking a side artificially when the strongest argument is something else.
What does a band 6 TOEFL Academic Discussion response look like?
A band 6 response takes a clear position in the first sentence, supports it with one specific example or reason, briefly acknowledges the strongest counter-argument, and resolves why the original position still stands. It uses varied sentence structures, contains no errors that obscure meaning, and references at least one of the two student posts by name to show genuine engagement.
Can I quote the professor's prompt or the other students directly in my TOEFL Academic Discussion?
Brief paraphrases are fine and signal engagement. Verbatim copying of long phrases is penalized as repetition and eats into your word count without earning content points. If you reference another student, use a short paraphrase plus the student's name. For example: "Like Paul argued, the bigger problem is enforcement, not the rule itself."
Can I use a personal example, or does it have to be academic?
Personal examples are not just allowed — they often score higher than abstract academic ones because they are concrete. A class you took, a job you held, a friend's experience, a news story you remember — all of these qualify. The key is specificity, not formality.
The Academic Discussion task is one of the most coachable on the new TOEFL Writing section. The framework is small, the rubric is clear, and most students who plateau do so because they are running an essay-format playbook on a forum-format task. Drill the four-step shape, force yourself to add one specific example per response, and always include the counter-and-resolve sentences. For deeper coverage of the other 2026 writing tasks, see our companion guides on the Write an Email task and the full Writing Section walkthrough.
Practice Academic Discussion with real 2026 prompts
Our free TOEFLMock writing tests include Academic Discussion prompts with expert evaluation and band feedback within 48 hours.
Start Free Writing Practice
DW
Daniel Whitaker
Head of Curriculum
Test preparation specialist and former classroom instructor. Designs full-length mock content aligned to the 2026 ETS redesign and writes study-plan, format, and score-requirement guides.