You analyze someone else's argument: one source, three short-answer questions. Worth 30% of the EOC. Each question is scored independently, so a weak Q1 will not drag down Q2. The suggested time here is 30 minutes, which protects your Part B time.
Q1 is 3 points, Q2 is 6, Q3 is 6. 15 points, each question scored on its own.
Part A is 30% of the End-of-Course Exam, and the exam is 45% of the whole thing.
Session 2.
On exam day, so you protect 90 for Part B.
In class you get three timed reps on a 60 / 50 / 45 ladder. On the real exam, about 30 minutes.
Your job is to identify, not to explain. Use the recipe: WHAT, WHY, SO WHAT. Capture all the parts of the argument, and only the parts the author actually states. One sentence is fine. College Board counts the parts.
Explain how, not just that. Walk the claims in order, and for every claim, say its job. The 4-versus-6 wall: listing the claims in order earns a 4. Explaining how each claim sets up the next, and how they all serve the main argument, earns the 6.
Evaluate, never just name. Naming tells the reader what a source is. Evaluating tells the reader what it can and cannot do for this specific claim. "The author cites a Stanford professor" is naming. It earns nothing on its own. Run two tests on every piece (credibility and relevance), work through at least three pieces, and name a real limitation that comes from the text.
WHAT: the author's main claim (the problem or the position).
WHY: the author's own "because," the reason or mechanism the author gives. This is the author's reason, not a study finding. Evidence is never a part of the argument; it supports the parts.
SO WHAT: what the author says should happen (the solution, the recommendation, or the implication the author states). When the author states one, it is a required part, not a bonus.
Sentence frame: The author argues that [WHAT] because [WHY], and therefore / so [SO WHAT].
Two checks before you move on:
1. Count your parts. Three present and accurate earns the 3. One missing, inaccurate, or swapped for evidence drops you to a 2. Order does not matter; presence and accuracy do.
2. Run the disagree test. Could a reasonable person disagree with what you wrote? If not, you wrote the topic ("the decline of handwriting"), not the argument ("schools should teach handwriting because...").
Parts are claims, never evidence. A study, a statistic, or a finding in the WHY slot means the WHY is missing. "Because a 2014 study found..." is support, not a part.
The actor matters. If the author says communities and states should invest, "parents should invest" is an inaccurate part. Name who the author says should act.
Keep the author's condition. "Leave phones alone when you are together" is the argument. "Minimize phone time" drops the condition and misstates the author.
Vague is not the same as missing. All three parts present, one of them fuzzy, still earns the 3. A part that is absent or wrong is a 2.
A restated title is a 1. Translate the argument into your own words. Readers name translation as the mark of understanding.
Lead with the spine. Put the WHAT first. It is the clearest way to write the sentence and it keeps you from burying the main claim. Order does not change your score: what is scored is whether every part is present and accurate.
Partial credit, in College Board's own terms: a restated title is a 1; the topic alone is a 1 or a 2; the claim without the author's recommendation, or the recommendation without the claim, is a 2.
For every claim: 1. Name the claim. 2. Say what it does for the argument. 3. Show what it makes possible. Then end with one sentence that ties the whole chain to the author's overall argument.
Job words (use these): sets up, builds on this by, backs it with, widens the problem by, raises the stakes, lets her end with, answers the objection that, narrows, depends on, makes the solution possible, establishes, defines, concedes, compounds.
Order words (banned in class responses when they stand alone): then, next, after that, finally, first, also, additionally. An order word is harmless when a job follows it ("Then, the author shifts to X, which does Y"). The ban trains the habit; the rubric scores the result.
The "In order to prove..." frame: In order to prove [main claim / thesis / argument], [author's last name] claims that [claim or reason]. To support this, [author's last name] uses [specific piece(s) of evidence]. Specifically, [author's last name]'s purpose in using this evidence is to [explain how it connects to the other claims and to the main argument]. Repeat for each major claim. The last sentence is where the points live: it is the "how," not the "what."
Naming the claim. Simple: The author claims that ______. Solid: [Author's last name] opens with the claim that ______. Sophisticated: [Author's last name] establishes early on that ______, a claim the rest of the argument depends on.
Saying its job. Simple: This claim sets up ______. Solid: This claim [backs it with / widens the problem by / raises the stakes by] ______. Sophisticated: By [establishing / defining / conceding] ______, the author makes it possible to argue that ______.
Showing what it makes possible. Simple: Because of this claim, the author can now say ______. Solid: Without this claim, the next claim about ______ would not be believable. Sophisticated: This claim does the groundwork for ______, which the argument could not reach otherwise.
Tying the chain (closing sentence). Simple: Together, these claims prove that ______. Solid: Each claim builds on the last, leading the reader to accept that ______. Sophisticated: The claims work in sequence: ______ makes ______ believable, which makes ______ matter, all building to the author's overall argument that ______.
Warning: a frame can be filled with pure summary. "This claim sets up ______" only earns credit when the blank is a job, not a restatement.
the move, the job words, the frames, and a two-claim practice.
1. The shuffle test. If your sentences would still make sense in a different order, you are labeling. Rewrite to show how one claim makes the next one possible.
2. The "without this claim" test. For each claim, can you finish the sentence "Without this claim, the author could not ______"? If you cannot, you have not found its job.
3. The destination test. Did you reach the last claim (the call to action, the recommendation, the thesis the author lands on)? Stopping before the destination holds at a 4 even when the connections you did write are strong. The setup claims count too: skipping the first third of the article holds at a 4.
Job paragraphs count. A paragraph with no claim and no evidence can still be the most important paragraph for Q2 when it defines the baseline the author's key term depends on. Read for the job of each paragraph, not only its content.
Evidence can appear in Q2, but it cannot replace the claim-to-claim connection. The 2025 readers named connecting claims to evidence rather than to other claims as a top Q2 mistake. Write the claim-to-claim link first; let the evidence follow.
From the Keys to Success handout.
Creating a foundation: introduces, establishes, defines, identifies, focuses, refers to.
Establishing the situation: analyzes, depicts, represents, speculates, generalizes, highlights, imagines, includes, presents.
Making a claim: argues, claims, advocates for, asserts, defends, implies, rejects, proposes, opposes.
Making distinctions: avoids, chooses, comments, compares, determines, develops, differentiates between, elicits, illustrates, indicates, maintains, manages, promotes, provides, qualifies, questions, rationalizes, regards, reveals.
Providing emphasis or support: alludes to, attributes, bases, demonstrates, details, discovers, discusses, displays, downplays, elevates, hints, interprets, justifies, overstates, outlines, recalls, recounts, selects, states, understates.
Evaluating: ascertains, assesses, assumes, attacks, challenges, empathizes with, forces, functions to, persuades, reasons, reflects on, verifies.
The borrow test: could the author have written this sentence without doing any research? YES: it is the author talking. That is a claim. NO: someone or something outside had to supply it. That is evidence. Number it.
Five signals to scan for: 1. Names: people, universities, organizations. 2. Numbers: statistics, percentages, years, amounts. 3. Quote marks: someone else is speaking. 4. Borrowing words: "a study found," "according to," "research suggests." 5. Specific examples: a real event, a real place, a real person's story.
One number per piece, not per sentence: a study and its statistic usually count as one piece of evidence doing one job. When two findings come from the same study, evaluate the study once.
Test 1, Credibility: Can I trust where this came from? Who is behind it? Could they actually see or measure this? What do they gain if you believe them? (This is RAVEN, applied. Never write the letters.)
Test 2, Relevance: Does it actually prove this claim? Right people, right size, right situation? Is there a gap between what the evidence shows and what the author says it shows?
A source can pass one test and wobble on the other. A brain scientist with a peer-reviewed study is credible; whether writing single words on a screen proves anything about taking real notes is a relevance question. That is why you run both tests on every piece. One test, or both done unevenly, is a 4. Both tests, every piece, is a 6.
Every piece of evidence has two parts: who said it and what they said. Running RAVEN on the person or institution is the credibility strand, and it is required. But that is half the question. The other half is what they said. Name the finding, say which of the author's claims it supports, and say one thing it does not prove. Who said it tells me whether to trust the source. What they said tells me whether the evidence actually does the job the author is using it for. The 2025 scoring guidelines are explicit: a response that evaluates only the sources, and not the specific evidence from those sources, cannot score a 6.
The per-piece frame: WHO (the source, and your credibility judgment with a reason) + WHAT (the finding, in your words) + FOR (which claim it supports) + BUT (one limitation of the finding, from the text) + VERDICT (fully, partially, or weakly supports, with a because).
Worked example (Gillis, on full-day kindergarten): WHO, a McMaster economist, credible on data but not an early-education researcher. WHAT, gains from full-day kindergarten are gone by the end of Grade 1. FOR, the claim that benefits are temporary. BUT, it is a quote to a magazine with no study or numbers behind it. VERDICT, partially supports, because the claim is exactly what Gillis needs but nothing in the article lets us check it.
Evidence Check frame (one per piece): While [this specific piece of evidence] [strength: why it is credible and relevant here], it [specific limitation], which means it [partially / strongly] supports the author's claim that [the specific claim it is attached to].
Open with one evaluative sentence about the evidence as a whole, then work through the pieces, then close by connecting back to the overall argument.
Three pieces is the standard. Two is College Board's floor. Every released 6 we can check evaluates three or more.
The limitation has to come from the text. "The economists might be biased because of taxes" is a guess; nothing in the article supports it. "The study is never named and we cannot check its sample" is a limitation you can point to.
Generic doubt and generic praise score the same. "The sources might be biased" earns nothing. So does "it comes from a reputable source."
Prestige is not credibility. "Stanford" is a name. "The reviewers are secondhand; they are describing someone else's study" is an evaluation.
Grouping is allowed; one pile is not. You may group similar evidence with one verdict per group, the way a released 6 grouped political scientists, real-world examples, and college professors. One pile with one verdict for everything is a label, and the 2025 readers describe that as "a singular, broad evaluation," a 4.
"Uneven" means one of two things: the detail is not consistent from piece to piece (real evaluation on one, perfunctory on the others), or one of the two strands is thin or absent. Both patterns hold at a 4.
Slips in attribution do not cap you when the evaluation is otherwise complete. Wrong evaluation does.
Write sentences a stranger could follow, not notes. The reader scores only what is on the page.
Useful words: credible, lacks credibility, effective, ineffective, strengthens, weakens, supports, relevant, lacks relevance, current, dated, representative, narrow, firsthand, secondhand, anecdotal, peer-reviewed.
Forces you through the whole article and maps every claim.
Evaluates the evidence you just mapped while it is fresh.
The fastest question on the exam once you know the argument cold. Run the recipe, count your parts, run the disagree test, done.
Exam-day order: answer Q2, then Q3, then Q1.
The topic or the title restated, or the argument misidentified.
Part of the argument: a part missing, overgeneralized, or inaccurate, or evidence standing in a slot.
Every part present and accurate, in your own words: WHAT, WHY, SO WHAT.
Claims listed with no connections, or the author's argument read backwards.
Claims named with order words or labels; connections acknowledged, not explained; the setup or the destination missing; evidence as the only link between claims.
Most of the claims, each with a job, the logic explained every time you move from one claim to the next, the destination reached and tied to the overall argument.
Evidence named or sorted into types; generic doubt or generic praise; fewer than two pieces.
Two pieces assessed thinly; credibility or relevance asserted instead of explained; sources judged but their findings never evaluated; one pile with one verdict.
Three or more pieces, each with credibility reasoned, relevance explained against the claim it supports, a limitation that comes from the text, and a verdict.
Answer a loaded EOC A passage (full or one question), then get AI practice feedback on the band you are near and the one move that would raise it. Practice feedback, not your real score.
What it looks like
The article is about the fading of penmanship.
What it should look like
Fink argues that schools should keep teaching handwriting, because handwriting practice makes writing automatic and frees a student's attention for what they are actually saying.
Run the disagree test. If nobody could disagree with your sentence, you wrote the topic.
What it looks like
Schools should keep handwriting because a 2007 study found handwritten essays were two years ahead of typed ones.
What it should look like
Schools should keep handwriting because practising it makes writing automatic, which frees a student's attention for the content.
The WHY is the author's own reason. A study supports a part; it is never a part.
What it looks like
Artushin argues that physical books beat screens because print improves comprehension and retention.
What it should look like
Artushin argues that physical books beat screens because print improves comprehension and retention, and therefore readers should keep choosing print, which also connects them to other readers in a way an e-reader cannot.
Check the last paragraphs. What the author says should happen is usually there, and it counts.
What it looks like
Gillis says parents should stop sending their children to full-day kindergarten.
What it should look like
Gillis says governments should stop funding universal full-day kindergarten and target the money at the children most at risk.
Name who the author says should act.
What it looks like
The author argues that we should let teenagers sleep.
What it should look like
The editors argue that schools should push start times to 8:30 or later, because teenagers' body clocks shift at puberty, and later starts improve health and academic performance.
Translate the argument into your own words. Title vocabulary is not understanding.
What it looks like
The author talks about e-readers, then about comprehension, then about bookstores.
What it should look like
Artushin opens with screens replacing paper so that the comprehension research lands as a cost we are already paying, not as a preference.
Ask what each paragraph is doing for the argument, not what it is about.
What it looks like
First she talks about sleep inertia. Next she talks about memory. Finally she talks about school.
What it should look like
She names sleep inertia first because it gives the problem a label, and that label is what lets her show the damage next.
Swap sequence words for connection verbs: establishes, builds on, complicates, reveals, concludes.
What it looks like
Then she claims that waking up suddenly causes sleep inertia.
What it should look like
Then she claims that waking up suddenly causes something called sleep inertia, and she supports this with how the brain works, since the decision-making part of the brain takes longer to wake up. This claim gives the problem a name, which lets her show the damage next.
Name the claim, say its job, show what it makes possible.
What it looks like
She claims print helps comprehension, and she uses a study that found comprehension is six to eight times better.
What it should look like
She claims print helps comprehension, which is what lets her next claim, that the body remembers pages, read as a mechanism rather than a feeling.
Write the claim-to-claim link first, then let the evidence follow behind it.
What it looks like
The author's last point is that the studies show gains fade by Grade 1.
What it should look like
The fading gains are not where Gillis stops: they set up his final claim, that the money should be targeted at the children most at risk instead of spent on everyone.
Ask what the last claim is doing at the end of the argument, then write it in with a job word.
What it looks like
The article starts by saying the provincial studies showed only small gains.
What it should look like
Before any of that, Gillis spends the opening on the government's promises for full-day kindergarten. That is the claim the whole argument exists to answer.
The setup claims have jobs too. Without the problem, your reasoning has a solution and nothing to solve.
What it looks like
The author is saying that schools should ban screens completely.
What it should look like
Artushin says readers should choose print, and that reading print connects them to other readers. She never calls for a ban.
Only claims the author actually makes.
What it looks like
The author's claim about technology addiction is not supported.
What it should look like
Artushin's words are “dopamine-addicted brains,” about scrolling, which is narrower than a claim about technology addiction in general.
Keep the author's terms. Renaming a claim means you evaluated your version of it.
What it looks like
The editors point to phones as the cause of teenagers' lost sleep.
What it should look like
The editors name phones only to set them aside, since “late nights are just part of the equation.” The cause they argue for is school start times.
Check the author's verdict on every example before you build the chain.
What it looks like
The 2007 study in the British Journal of Educational Psychology is credible, so the evidence is strong.
What it should look like
The 2007 study was published in the British Journal of Educational Psychology, which makes it credible, but it is fifteen years old, and what a student can do with a keyboard has changed since then.
Name a specific weakness in at least one source, and say why it matters for the claim it supports.
What it looks like
The author quotes a McMaster economist, and McMaster is a respected university.
What it should look like
WHO, a McMaster economist, credible on data but not an early-education researcher. WHAT, gains from full-day kindergarten are gone by the end of Grade 1. FOR, the claim that benefits are temporary. BUT, it is a quote to a magazine with no study or numbers behind it. VERDICT, partially supports, because the claim is exactly what Gillis needs but nothing in the article lets us check it.
You did the who. Now add the what.
What it looks like
The author uses studies, experts, and examples, and they are all credible.
What it should look like
The provincial studies and the California study both measure outcomes, so together they carry the claim that gains fade; the quote to a magazine cannot, because no numbers come with it.
Group the evidence by what each group proves, and give each group its own verdict.
What it looks like
The economists are probably biased because they do not want to pay higher taxes.
What it should look like
The economist's claim reaches Gillis as a quote to a magazine, with no study and no numbers behind it.
A limitation has to come from the text. Point to it.
What it looks like
Rothkopf's research is credible because it was published.
What it should look like
Rothkopf (1971) is credible on memory, but his work predates e-readers, so it cannot speak to print against screen.
Say why that expertise matters for this particular claim, and where it runs out.
What it looks like
Two full evaluations, and then: “Also the third source is credible.”
What it should look like
Two full evaluations, and then a third in three sentences: who the source is, what it found, and how well it holds up the claim.
Three pieces, each shorter than the last. Timing, not skill.
You found it. Ten questions about this course, all answerable from this website. There are three versions of the gauntlet and you get a random one each time. A perfect 10 for 10 unlocks a 3-minute BRAIN BREAK that is definitely not schoolwork. When the 3 minutes are up, the arcade closes and you have to beat a fresh version to reopen it.
Arrow keys to move, Space (or tap) to shoot. Five themed waves of point-killers are descending on your Works Cited: weak sources, thesis problems, evidence sins, reasoning gaps, and a final boss wave of academic integrity violations. Three lives, and the whole arcade stays open for exactly 3 minutes. When the break clock hits zero, it is back to the books.
IRR Purpose: Explore the complexity of an issue by examining it through different perspectives and connecting it to the larger context.
Expert 1 (with credentials)
Expert 2 (with credentials)
Expert 3 (with credentials)
Expert 1 (with credentials)
Expert 2 (with credentials)
Expert 3 (with credentials)
Expert 1 (with credentials)
Expert 2 (with credentials)
Expert 3 (with credentials)
Save the PDF, then upload it to the Canvas assignment to turn it in to Mrs. Cohen.
Expert 1 (with credentials)
Expert 2 (with credentials)
Expert 3 (with credentials)
RAVEN is how this course evaluates whether a source deserves your trust. Run every key source through all five letters, then decide: Strong, OK, or Weak. The interactive Source Tracker (in Templates) applies this to your full source list.
What is the reputation of the author, source, and publication? Is the publication peer-reviewed? Do past actions or statements suggest unreliability? Is the author or publication in a position of authority on this topic?
Is the author in a position that allows access to reliable evidence? If the piece is about an event, did the author actually observe it firsthand?
Does the author have a personal stake in the topic? Would they gain anything by lying? Would they gain anything by telling the truth?
Does the author have specialized knowledge on the topic? Does the evidence itself come from a source with real expertise?
Is the author neutral, or is bias evident? If there is bias, is it evidence-based? Is the source of the evidence itself neutral?
On the exam and in your papers: never name RAVEN. Apply it. "M. Stacia Dearmin, a practicing physician at Akron Children's Hospital, writes from direct clinical experience" is Expertise and Ability to Observe in action, without the acronym.
If the tracker does not appear, reload the page.
A 6 on Q2 names each claim and says what it does for the next one. Order words are labels; jobs are explanations.
"This claim [job], which builds the author's central argument." A top-scoring 2025 paper wrote that sentence ten times. The readers cared only that every claim had a job.
Order words (then, next, finally) are fine only when a job follows them.
A paragraph with no claim can still have a job, when it defines the baseline the argument depends on.
Verbs: establishes, introduces, defines, sets up, identifies.
Stem: Claim ___ establishes ___, which the argument needs in place before it can ___.
Verbs: explains why, backs it with, depends on, gives the problem a name.
Stem: Claim ___ builds on the one before it by taking ___ and pushing it toward ___.
Verbs: widens the problem by, raises the stakes, compounds, makes the problem harder to brush off.
Stem: Claim ___ reveals the stakes: if ___, then ___.
Verbs: concedes, answers the objection that, qualifies, refutes.
Stem: Claim ___ complicates things by showing ___ is not as simple as it seemed, because ___.
Verbs: makes the solution possible, lets her end with, concludes, reaches.
Stem: Because the earlier claims showed ___, the argument concludes that ___.
Naming the claimThe author claims that ______.[Author] opens with the claim that ______.[Author] establishes early on that ______, a claim the rest of the argument depends on.
Saying its jobThis claim sets up ______.This claim [backs it with / widens the problem by / raises the stakes by] ______.By [establishing / defining / conceding] ______, the author makes it possible to argue that ______.
What it makes possibleBecause of this claim, the author can now say ______.Without this claim, the next claim about ______ would not be believable.This claim does the groundwork for ______, which the argument could not reach otherwise.
Tying the chainTogether, these claims prove that ______.Each claim builds on the last, leading the reader to accept that ______.The claims work in sequence: ______ makes ______ believable, which makes ______ matter, all building to the author's overall argument that ______.
In the IRR[Name] argues ______; that claim rests on ______.
Two claims that sit next to each other in Maria Konnikova's “Snoozers Are, in Fact, Losers,” a passage the class has worked. Write the sentence that links them.
Claim 1. Snoozing does not give you more rest, and it makes waking up harder.
Claim 2. Waking up suddenly causes something called sleep inertia.
A model link: Then she claims that waking up suddenly causes something called sleep inertia, and she supports this with how the brain works, since the decision-making part of the brain takes longer to wake up. This claim gives the problem a name, which lets her show the damage next.
Notice the job: the second claim does not just come after the first, it names the problem the first one described, and naming it is what lets Konnikova measure the damage in the claim after that. That is the move the 6 is scored on.
Five tests, and one tie-breaker question underneath. Type your question, enter the class code, and press Check my question. You get three checks a day, with a short wait between them, so revise before you check again.
Cannot be answered with yes, no, or one word.
Not yet looks like: "Is fast fashion bad?" "How many teens use social media?"
The fix: a stem that forces a judgment. To what extent should or could... What role does [factor] play in... How might [stakeholder] address... What responsibility does [group] have to... How effective is [approach] at...
Names a particular group, place, time, or practice. College Board's own scoring example: “water pollution in India,” not “pollution.”
Not yet looks like: "What are the effects of pollution?" Air, water, and land is three papers.
The fix: the specific-noun move. Swap every general noun for the specific one you actually care about. “Teenagers” becomes “high school athletes.”
Reasonable people, with evidence, would answer it differently. A topic with no position cannot make an argument.
Not yet looks like: "How does X help Y?" It has already decided. "How does air pollution affect children?" Nobody argues the other side.
The fix: the tension move. Two real things pull against each other; somebody has to give something up. Name the stakeholder who loses if your answer wins.
At least half of the articles on your table touch it, and you could finish by October 7.
Not yet looks like: "How does social media affect teens, and what should schools do about it?"
The fix: split it and keep the one you can finish. Then check the published conversation: what are the writers in your articles already asking?
Someone besides you would want to read 1,200 words about it. Interesting is a skill, not luck.
Not yet looks like: a question a smart stranger answers with "sure, obviously."
The fix: the only-you detail (one thing you know from your life or your community that most adults would not think to ask) and the stranger test (a smart stranger says “huh, I never thought about that”).
The same checks, grouped the way the class uses them.
Pick which task you are writing for. The two are different on purpose: the IWA takes a side; the IRR frames a conversation without taking one. Your work saves automatically on this device.
Remember: the IRR reports the conversation. Your framing statement promises perspectives and analysis, never a side. Save your opinion for the optional recommendation at the very end.
Then run it through the Five Key Qualities above. If you cannot answer yes to all five, revise before you draft.
A thesis is not just a topic or opinion. It is your main argument, the thing you will prove using evidence. Ask: What do I want to argue? What do I want my reader to believe by the end?
Sum up a source in four sentences so you understand it before you use it. (Teachers call this a rhetorical précis.)
[Author's name], [credentials], in [his/her/their] [genre], [title] ([publication info]), [rhetorical verb] that [main claim].
Rhetorical verbs. Tap one to drop it into your sentence.
[Author] supports this claim by [method/approach], revealing that [key findings] ([citation]).
Types of support to look for. Tap one to drop it in.
[Author]'s purpose is to [goal] in order to [intended outcome or effect on audience].
Purpose verbs. Tap one to drop it in.
Given [author]'s findings about [key point], this work supports my argument that [your claim].
A source might provide evidence for one of your claims, establish context or significance, offer a perspective you build on, or represent a counterargument you will address.
Given [author]'s findings about ______, this source [supports / complicates / rebuts] the perspective that ______.
The IRR has no "my argument," so this sentence places the source among the perspectives. Tap the move it makes.
Sentence frames for AP Seminar, grouped by the move they make and the rubric row that move earns. Pick a move, then a tier. Start at Simple, and move up when the simple version stops saying everything you mean.
IRR Row 4, IMP Row 3
A perspective is a source's argument, not a lens and not a group of people. Name who argues what, and why.
[Name], a [credential], argues that ______.
[Name], a [credential] writing in [publication], argues that ______ because ______.
Writing from [position or stake], [Name] argues that ______, a claim that rests on ______ and that [another source] will [complicate / contest] below.
IRR Row 3, EOC A Q3
Say why this person or study can be trusted on this claim; never write the letters.
[Name], who [credential or firsthand position], reports that ______.
Because [Name] [observed / studied / measured] ______ directly, the finding that ______ carries weight on the question of ______.
Although [Name] has [a stake in / a distance from] the outcome, the [peer-reviewed / government / firsthand] basis of the finding that ______ makes it [strong / usable with care] evidence for ______.
IRR Row 4, IWA, EOC B Row 3
Sources rebut, complement, support, limit, or constrain one another. The test: remove one source and the paragraph should break.
To extend[Source B] agrees, adding that ______.
To extendTo further support [Source A]'s finding that ______, [Source B] states ______.
To extend[Source B]'s findings help explain why [Source A] observed ______: ______.
To disagreeHowever, [Source B] disagrees, claiming that ______.
To disagree[Source B] rejects [Source A]'s claim that ______ on the grounds that ______.
To disagree[Source A]'s claim that ______ rests on the assumption that ______, which [Source B]'s evidence of ______ undercuts.
To complicate[Source C] adds a complication: ______.
To complicateWhere [Source A] emphasizes ______, [Source C] complicates this by ______.
To complicateAlthough [Source A] is partially right about ______, [Source C]'s evidence that ______ suggests the picture is ______.
To name the splitThe two sources disagree about ______.
To name the splitThese perspectives agree that ______ but split over ______.
To name the splitThese perspectives agree that ______ but split sharply over ______, and that disagreement matters because ______.
IRR Row 2, EOC A Q2
Give every claim a job.
[Name] claims that ______. This claim sets up ______.
[Name] opens with the claim that ______, which [backs it with / widens the problem by / raises the stakes by] ______.
By [establishing / defining / conceding] ______, [Name] makes it possible to argue that ______, which the argument could not reach otherwise.
IRR Row 3, EOC A Q3 relevance
Say what the finding proves for the claim it sits next to, and what it does not.
This evidence shows ______, which supports the claim that ______.
The finding that ______ supports [Name]'s claim that ______ because ______; it does not prove ______.
While [this finding] [strength], it [specific limitation], which means it [partially / strongly] supports the claim that ______.
IRR Row 4 and the conclusion of everything
Together, these sources show that ______.
Taken together, these sources suggest ______, a conclusion none of them reaches alone.
If [Perspective A] gets their way, ______; if [Perspective B] gets their way, ______. Either path means ______, which is why ______.
IRR Row 1 and Row 3, the oral defense
One limit of this research is ______.
[Source] cannot account for ______, because its [sample / scope / date] ______.
What none of these sources measures is ______, which leaves the question of ______ open and is where further research would need to begin.
Structured Student Talk
Use these out loud before you write. Paraphrase first, then build.
No tiers for this one.
ParaphraseSo what [classmate] is saying is ______. / I heard [classmate] say ______. / Big ideas presented include ______. / Essentially, ______.
BuildI agree with the idea that ______ because ______. / I disagree with the idea that ______ because ______. / Adding on, ______. / This connects to ______. / Something I noticed is ______. / This makes me wonder ______. / That resonates because ______. / A question I had is ______.
IWA and EOC B only; in the IRR, only in the optional recommendation
The evidence ultimately supports ______.
While both sides make valid points, the evidence ultimately supports ______ because ______.
While [Source A] argues ______ and [Source B] emphasizes ______, I argue ______ because ______.
Derived from Graff and Birkenstein's They Say, I Say. Think of these as moves in a sport: the better you get, the more you know when and how to use each one.
Adding: moreover, furthermore, in addition, indeed · Contrasting: however, nevertheless, conversely, on the other hand · Comparing: likewise, similarly, along the same lines · Examples: for instance, specifically, as an illustration · Results: consequently, as a result, therefore · Conceding: admittedly, granted, to be sure · Concluding: in short, all in all, thus
Directions: Peer review each teammate's IRR using the AP Seminar rubric. Complete a review table for each person in your group (you do not review yourself).
Why: This helps you understand the rubric on a deeper level so you can produce the best possible IRR, you help your teammates, and afterward you have read every teammate's IRR, a necessary step for the TMP.
When finished: submit on Canvas and email the review to your teammates so they can view their feedback.
Check each box as your team finishes it. Keep this on top of your deck. Run time goal: 8 to 10 minutes.
Lock in the speaker order: who opens, who closes, and every handoff line. For each part, lock in: main point in one sentence, the evidence used (author/title or site), and visuals. Check before rehearsal: every slide shows short text plus credit, no repeated points across speakers, and total time adds up to 8 to 10 minutes.
Speaker order, from who opens to who closes: Opener/Context, Solution leads, Limits/Counter, Conclusion/Implications. Support roles: Slide driver, Timekeeper, and Evidence verifier (checks on-slide credits).
Run this top to bottom before presentation day.
Eye contact and energy are the two most heavily scored components. Also: vocal variety, movement, not over-relying on note cards, and cards that are easy to handle.
Since this is an argument paper, you must: present a debate with multiple perspectives, answer your research question, and propose a solution or resolution. 2,000 words max (excluding citations), at least one stimulus source, scholarly sources, and MLA or APA citations with a Works Cited.
Write your research question, your working thesis (your answer to the question), and the stimulus source(s) you will use with a note on how each one enters your argument.
For each section: a main claim, the sources and key evidence you will use, and your analysis explaining how it supports your thesis. Put sources in conversation rather than summarizing each one separately, connect every source back to your question and thesis, address multiple perspectives, and add your analysis: you are the expert now.
Required for top scores. What would someone who disagrees say, and how do you respond? You do not always need to refute: you can confirm, concede, qualify, adjust, or redefine.
Your solution or resolution (required), its limitations (is it applicable everywhere, for all people?), and its implications (what are the consequences if it is adopted?).
Four more outline formats (Classic Argument Model, claim-and-evidence grids, the paragraph-by-paragraph I-E-E-C template, and the rubric-aligned planner) are in the class IWA Outline doc if this structure does not fit your topic.
One source to analyze. Three questions, each scored independently. In class, you get three timed practice rounds, and they step down as your speed builds: 60 minutes for your first, then 50, then 45. On the real exam, budget about 30 minutes for Part A so you protect 90 for Part B.
Write Part B first. Then come back to Part A and answer in this order: Q2 (line of reasoning), then Q3 (evidence), then Q1 (argument). Q2 forces you through the whole article and maps every claim; Q3 evaluates the evidence you just mapped while it is fresh; Q1 is the fastest question on the exam once you know the argument cold.
"Source X has Expertise" earns a 4. "Source X's expertise in Y allows the author to support the claim that Z" earns the 6. Same rule on Q2: naming a connection earns a 4; showing how one claim makes the next one possible earns the 6.
A mini IWA: 4 sources connected by a theme, no more than 90 minutes. The College Board sources are the backbone of your evidence; you only need to incorporate and synthesize 2.
Although (counterargument), (main argument) because (supporting 1 and supporting 2).
Okay, but it should not carry your argument: it tends to read as anecdotal. Its best home is your hook or conclusion.
Row 1 Establish Argument, Row 2 Line of Reasoning, Row 3 Select and Use Evidence, Row 4 Apply Conventions: each 0/2/4/6. A response can be a 6 on Row 1 and a 2 on Row 3.
Every term below shows up in the rubrics, the prompts, or Mrs. Cohen's feedback. The Big Six come straight from the classroom Key Definitions poster, with the same examples.
| Term | Definition | Example |
|---|---|---|
| Lens | A focus point through which to view an issue. | Looking at the political impacts of climate change compared to the scientific impacts versus the social impacts. |
| Perspective | A specific argument or point of view about an issue. The test: can someone disagree with it? Scored in: IRR Row 4, the IWA, and EOC B, wherever you put arguments in conversation. | The most important step on climate change is holding major corporations responsible vs. implementing more climate-friendly legislation. |
| Stakeholder | An individual or group affected by the ideas and decisions around an issue. | Government officials, scientists, researchers, consumers, CEOs of major corporations. |
| Term | Definition | Example |
|---|---|---|
| Implication | A possible future effect or result. | Increased climate-friendly legislation may lead to more accountability for climate-friendly practices. |
| Argument | A claim or thesis that conveys a perspective, developed through a line of reasoning and supported by evidence. Scored in: EOC B Row 1 and IWA Row 1 (Establish Argument). IRR Row 2 is where you analyze the arguments inside your sources. | For consumers to hold companies responsible, they must make their stances clear through purchasing habits and collaborate with ethical companies to create real social change. |
| Line of Reasoning | The purposeful arrangement of claims with supporting evidence that leads to a conclusion. Scored in: EOC B Row 2 (Line of Reasoning), IRR Row 2, EOC A Question 2. | Each claim sets up the next; shuffle the paragraphs and the argument should break. |
The difference between a list and a line of reasoning is whether the order matters. Shuffle each set and watch what happens.
| Term | What it means in this course |
|---|---|
| Claim | A single debatable statement your evidence must prove. Topic sentences carry claims; your thesis is your biggest claim. Scored in: EOC B Rows 1 and 2; every topic sentence in the IWA and IRR. |
| Evidence | Facts, statistics, expert findings, examples, or testimony used to support a claim. Evidence never speaks for itself. Scored in: EOC B Row 3 (Select and Use Evidence), IRR Row 3, EOC A Question 3. |
| Commentary | Your explanation of what the evidence proves and how it supports the claim. This is where rubric points actually live. Scored in: EOC B Row 3 and IRR Row 2: the reasoning right after each citation. |
| Counterargument | The strongest opposing view, presented fairly, not a strawman version of it. Scored in: EOC B Row 1: a complex argument engages the strongest opposing view instead of dodging it. |
| Rebuttal / Refutation | Your response showing why your argument holds despite the counterargument. |
| Concession | Honestly granting the part of the counterargument that is true before explaining why your position still stands. |
| Qualification | Limiting a claim so it stays defensible: "in most urban districts" instead of "everywhere, always." |
| Call to action | What the author wants people to do, change, support, or stop. Finding it is the fastest route to an author's argument on EOC A Q1. |
| Term | What it means in this course |
|---|---|
| Credibility | Whether a source deserves trust, judged with RAVEN: Reputation, Ability to observe, Vested interest, Expertise, Neutrality. Scored in: EOC A Question 3, IRR Row 3, IWA Row 3. Apply RAVEN; never name it. |
| Relevance | Whether the evidence actually bears on the specific claim it sits next to. Credible evidence can still be irrelevant. |
| Validity | Whether reasoning or a study design actually measures or proves what it claims to. |
| Reliability | Whether a source or method produces consistent, repeatable results. |
| Bias | A leaning that shapes how information is selected or framed. Every source has some; your job is to name how it matters. |
| Vested interest | When a source stands to gain (money, power, reputation) from a particular conclusion. |
| Limitation | A specific weakness of evidence: small sample size, missing credentials, dated data, vested interest, or narrow scope. A limitation has to come from the text. Guessing that a source might be biased earns nothing. |
| Job word | A verb or phrase that names what a claim does for the argument (sets up, builds on this by, backs it with, widens the problem by, raises the stakes). The opposite of an order word. |
| Order word | then, next, after that, finally, first. Describes sequence, not reasoning. Harmless when a job follows it, a 4 when it stands alone. |
| Load-bearing background | A paragraph with no claim and no evidence that still does a job for the argument, usually defining a baseline so a key term makes sense. Read for the job, not only the content. |
| The borrow test | Could the author have written this sentence without doing any research? If yes, it is a claim. If no, it is evidence. |
| Subject, Topic, Issue | The ladder you climb down to find a research question. A subject is too big to research; an issue is a specific problem with stakeholders and a real tension. |
| Scholarly / peer-reviewed | Work checked by other experts before publication. The backbone of a Tier 1 source list. |
| Term | What it means in this course |
|---|---|
| Context | The who, what, where, when, and why-it-matters that situates your issue in a larger conversation. IRR Row 1 territory. |
| Complexity | Treating an issue as having multiple causes, stakeholders, and trade-offs rather than one obvious answer. |
| Synthesis | Putting sources in conversation until they say something together that none says alone. The most rewarded move in the course. |
| Stimulus material | The College Board packet of texts released in January. Your IWA must connect to a theme from at least one of them, organically. |
| Resolution / Solution | The realistic answer your argument lands on, always presented with its limitations and implications. |
| Plagiarism / Attribution | Using ideas or words without credit vs. clearly naming where every idea came from. Missing attribution can zero out an entire task. |
Want these in flashcard form? Cover the right column, read the left, and define out loud. The terms repeat across every assessment, so this list works for the EOC too.
The test: can a reasonable person argue against it? No: topic, lens, or stakeholder. Yes: perspective.
The sunglasses analogy. A lens works like a pair of tinted sunglasses: it filters what you see. Put on the economic lens and money questions jump out; put on the ethical lens and right-and-wrong questions jump out. A lens is a way of looking. A perspective is what someone argues once they have looked.
Ask: can a reasonable person argue against it? If no, you have a topic, a lens, or a stakeholder. If yes, you have a perspective. "Scientists" is not a perspective. "Scientists who argue geoengineering is too risky to deploy" carries a perspective.
One lens, many perspectives Inside a single lens you should still find 2 to 3 differing perspectives. The economic lens on housing contains economists who blame zoning, economists who blame interest rates, and developers who blame construction costs. The lens is shared; the arguments clash.
| Level | What it is | Example | Can someone disagree? |
|---|---|---|---|
| Topic | The issue itself | Climate change | No. It is just a subject. |
| Lens | A focus point for viewing the issue | The economic lens on climate change | No. It is just a viewing angle. |
| Stakeholder | A person or group affected by the issue | Fossil fuel workers; coastal homeowners | No. They are people, not claims. |
| Perspective | A point of view conveyed through an argument | "Carbon taxes are the most effective climate policy because they price the true cost of emissions" | Yes. This is the test. |
IRR Row 4 and IMP Row 3 both score perspectives, not lenses. A paper that says "the economic perspective and the political perspective" without naming what anyone is actually arguing reads as oversimplified, the 2-point descriptor. Name WHO argues WHAT and WHY, then connect those arguments to each other.
"While [Perspective A] assumes ___ in order to achieve ___, [Perspective B] counters that ___, which suggests ___."
"Economists like [name] argue ___, but public health researchers such as [name] complicate this by showing ___."
"These perspectives agree that ___ but split sharply over ___, and that disagreement matters because ___."
What synthesis is: putting two or more sources in conversation inside the same paragraph until they say something together that neither says alone. What it is not: summarizing Source A, then summarizing Source B, then moving on. Where it scores: IRR Row 4, IWA evidence, EOC B Row 3, and the conclusion of everything.
1. Clarify: use one source to sharpen or explain another. The employers' 7.7 out of 10 shows the pattern across the trials; the Irish employee's calendar shows what it looks like in one week.
2. Complicate: use one source to show what another misses. Where one source emphasizes a gain, another complicates it by documenting a cost the first never measures.
3. Push back: use one source to directly challenge another, then say who has the stronger case and why. Your commentary referees the disagreement.
What it looks like
The UK trial of 73 companies and 3,300 employees found that almost half of respondents said productivity improved slightly or significantly (WEF). Commuting time per week fell from 3.5 hours to just under 2.6 hours, 27% lower (WEF). Economist Juliet Schor argues that a shorter working week is key to cutting global carbon emissions (WEF). These sources show different views on the four-day week.
What it should look like
The UK trial's productivity result seems to settle the question, but it comes from the workers reporting on themselves: almost half said productivity improved, and 86% wanted to keep the schedule, which measures how the week felt as much as what it produced. The commuting finding is the harder number, a drop from 3.5 hours a week to just under 2.6, and it is what lets Schor's argument about carbon emissions attach to this trial at all. Together, the three suggest the strongest case for the four-day week may not be the productivity claim the article leads with but the hours nobody spent commuting, a case the article never makes for itself, and one it never tests against anyone who lost out.
Notice what changed: the sources now react to each other (seems to settle, is the harder number, lets Schor's argument attach), the writer's commentary connects them, and the paragraph ends somewhere none of the sources went alone. Every number is the article's own.
Each one opens in Stuck on a sentence? at the conversation frames, in all three tiers.
The College Board AI rule on synthesis. Asking AI to compare, contrast, or synthesize across sources has no acceptable use. Synthesis is the skill the course assesses; checkpoints and the oral defense verify you did it yourself.
For more frames, open For a model, open any top-scoring IRR sample and watch the "if A gets their way / if B gets their way" move in its conclusion.
Exploring possible topics of inquiry; a first search for sources; confirming your understanding of a complex text; checking your writing for grammar and tone.
You read primary and secondary sources directly; you do your own analysis and synthesis of the evidence; you make your own choices about how to communicate, in writing and in presentations; you engage deeply with credible, valid sources and bring in diverse perspectives.
A team, or a single team member, that does not follow the College Board policy on generative AI for the Team Multimedia Presentation receives a group score of 0 for that component. One person's choice costs the whole team.
The same boundaries apply to the practice IRR, the team slide, and Current Events. College Board does not allow AI-made slides, so neither does this class.
using AI to get a sense of existing debates, sub-topics, or what is widely known.
taking AI output uncritically, like having AI generate your research question or thesis, or relying on AI as your source of information.
asking AI for authors, organizations, or publications to then locate and read directly, and asking for recommendations to fill gaps.
using an AI-generated source list without going to the original sources. AI invents sources; verify every one is real, credible, and relevant.
help with complex vocabulary, sentence structures, or a confusing passage.
generating a summary instead of reading, requesting quotes or citations you did not find yourself, or pasting AI summaries into a draft. Always read the original text.
Asking AI to compare or contrast sources, synthesize across sources, or develop statements that put sources in conversation is prohibited. This is the skill the course assesses; checkpoints will ask you questions to ensure you did this work yourself.
general best practices for structuring a paper.
AI-produced outlines or drafts of your specific paper, or any AI-generated writing in the final draft.
spell and grammar checkers, and feedback on style and tone where you deliberately choose what to incorporate.
accepting AI revision suggestions uncritically or inserting new AI-suggested sections.
guidance on citation format, and drafting or checking the format of a bibliography for sources you actually studied.
AI citations for sources you have not read, or trusting AI formatting without checking it.
general guidance on effective presentations and initial ideas for key points, sequence, or visuals.
uncritically using AI to produce your key points, visuals, or structure, or memorizing an AI-generated script.
Do not use AI to generate possible answers to defense questions.
Curve formula: Curved % = square root of (raw / max) x 100, floored to the nearest whole number. The curve always rounds down, so what you see is exactly what you receive.
| Raw | Raw % | Curved % | Grade |
|---|---|---|---|
| 15 | 100% | 100% | A |
| 14 | 93% | 96% | A |
| 13 | 86% | 93% | A |
| 12 | 80% | 89% | B |
| 11 | 73% | 85% | B |
| 10 | 66% | 81% | B |
| 9 | 60% | 77% | C |
| 8 | 53% | 73% | C |
| 7 | 46% | 68% | D |
| 6 | 40% | 63% | D |
| 5 | 33% | 57% | F |
| Raw | Raw % | Curved % | Grade |
|---|---|---|---|
| 24 | 100% | 100% | A |
| 22 | 91% | 95% | A |
| 20 | 83% | 91% | A |
| 18 | 75% | 86% | B |
| 16 | 66% | 81% | B |
| 14 | 58% | 76% | C |
| 12 | 50% | 70% | C |
| 10 | 41% | 64% | D |
| 8 | 33% | 57% | F |
| Raw | Raw % | Curved % | Grade |
|---|---|---|---|
| 30 | 100% | 100% | A |
| 28 | 93% | 96% | A |
| 26 | 86% | 93% | A |
| 24 | 80% | 89% | B |
| 22 | 73% | 85% | B |
| 20 | 66% | 81% | B |
| 18 | 60% | 77% | C |
| 16 | 53% | 73% | C |
| 14 | 46% | 68% | D |
| 12 | 40% | 63% | D |
| 10 | 33% | 57% | F |
Practice IRR. The real IRR is scored by College Board readers and does not receive a class grade.
| Raw | Raw % | Curved % | Grade |
|---|---|---|---|
| 18 | 100% | 100% | A |
| 16 | 88% | 94% | A |
| 14 | 77% | 88% | B |
| 12 | 66% | 81% | B |
| 10 | 55% | 74% | C |
| 8 | 44% | 66% | D |
| 6 | 33% | 57% | F |
One piece of evidence, five moves. Three pieces, fully done, is the 6.
Five moves, one worked example each, from Artushin on Mangen's Kindle study.
Is this source trustworthy on this claim, and why? (Reputation, ability to observe, vested interest, expertise, neutrality, applied, never named.)
“Mangen studies reading, and she compared paper and Kindle readers herself, so she saw the difference directly.”
What did the evidence actually find?
“Paperback readers put the plot events in order better than Kindle readers.”
Which of the author's claims does it support?
“Artushin's claim that paper improves comprehension.”
What limitation comes from the text?
“One story, a small group, and the article gives no numbers.”
Fully, partially, or weakly supports, with a because.
“Partially supports, because it shows better sequencing, not better understanding.”
One row per piece of evidence, five boxes per row. Saved on this device. Then see your evaluation as a paragraph before you write the real one.
Piece 1
Piece 2
Piece 3
A limitation has to point to a sentence in the passage. A guessed one scores nothing.
The Q3 row of “What each score looks like.”
Evidence named or sorted into types; generic doubt or generic praise; fewer than two pieces.
Two pieces assessed thinly; credibility or relevance asserted instead of explained; sources judged but their findings never evaluated; one pile with one verdict.
Three or more pieces, each with credibility reasoned, relevance explained against the claim it supports, a limitation that comes from the text, and a verdict.
Three pieces fully done is the standard. Two is College Board's floor. Ten one-liners is a 4.
Sources only, no findings evaluated, cannot score a 6.