Skip to content

AP Psychology · technique

How to answer the free-response section

On this page · 5 sections

Section II is 70 minutes, two questions, one third of your score. There is no essay in it.

Question 1 is the Article Analysis Question (AAQ). You read a short summary of one research study — method, participants, procedure, results — and answer six prompts labeled A through F. It is worth 7 points, and the exam suggests 25 minutes. Parts A through E are worth 1 point each. Part F is worth up to 2. The six prompts are predictable. One asks you to name the research method. One asks how a variable was measured, which is its operational definition. One asks what a reported statistic shows. One asks for an ethical guideline the researchers followed. One asks how far the findings reach beyond these participants. The last asks you to use a result to argue for or against the hypothesis, or a psychological concept the question names.

Question 2 is the Evidence-Based Question (EBQ). You get three summarized studies labeled Source 1, Source 2, and Source 3. You take a position on a question and defend it using evidence from at least two of them. It is also worth 7 points, and the exam suggests 45 minutes. Its parts are A, B(i), B(ii), C(i), and C(ii). Part A is worth 1 point, Part B(i) 1, Part B(ii) 2, Part C(i) 1, and Part C(ii) 2. Section II is 14 raw points in total.

Watch the labels, because the two questions label things differently. On the AAQ your answers are A, B, C, D, E, F, with no roman numerals anywhere. On the EBQ your answers are A, B(i), B(ii), C(i), C(ii). The three EBQ sources are numbered, never lettered. Every piece of evidence you use has to point at its source. Two forms are unmistakable: a parenthetical such as (Source 2) at the end of the sentence, or a lead-in that opens with According to Source 2. Other clear citation styles are accepted, so do not lose time over format. Evidence with no citation at all does not earn the evidence point. Calling it Source B when you mean Source 2 blurs the one thing those rows require you to name.

Both questions are scored point by point against a rubric, not read as writing. Nothing is awarded for an introduction, a conclusion, elegant phrasing, or length. Most rows are 1-point and all-or-nothing: a reader either finds the task completed or does not. Three rows carry partial credit. AAQ Part F, EBQ Part B(ii), and EBQ Part C(ii) are each scored 0 to 2. Those three rows hold 6 of the section's 14 points, so plan to spend real time on them.

Two habits protect points everywhere in this section. Answer in complete sentences, because bulleted fragments and bare definitions earn nothing. And keep every sentence accurate. Inaccuracy is punished three different ways, so learn the ledger. On a 1-point row, an inaccurate claim sitting next to the correct one can cost you the point. On AAQ Part F, misreading the result caps you at 1 of 2. On EBQ Part B(ii) and Part C(ii), the explanation has to rest on an accurate reading of the evidence. Misread the source and you can lose the explanation credit, however well you apply the concept, so no extra clause is worth the risk.

The steps

  1. 1

    Split the clock before you write a word

    Twenty-five minutes on the AAQ, forty-five on the EBQ. Those two totals are the exam's own suggested times. The split inside each is our recommendation: on the AAQ, about 10 minutes reading the study and planning and about 15 writing. On the EBQ, spend about 15 minutes reading the three sources and planning and about 30 writing. The split is deliberately generous on reading because both questions punish a misread stimulus more than a plain sentence. Do the AAQ first. Its points are mechanical — name the method, quote the measurement, read the numbers — so you can bank them at a steady rate. Set a hard stop at minute 25. If a subpart is not done, leave it and move, because an unwritten EBQ claim costs more than one blank AAQ part. Every minute you spend past 25 comes straight out of EBQ writing time, and the harder points live there.

  2. 2

    Read the prompts before the passage

    The prompts tell you what to hunt for, so reading the study first means reading it twice. Go to the questions first and pick out the task verb and the exact noun attached to it. Then, as you read the study, use the Bluebook highlighter on the sentences that answer them. If Part B asks for the operational definition of aggression, you are hunting one sentence that says how aggression was counted — not the whole procedure. If Part E asks about generalizability, you are hunting the participants sentence. On the EBQ, read the question, then skim each source for its finding and its method, and jot a two-word note on each on your scratch paper. You will build Part B around one source and Part C around another; the notes tell you which two.

  3. 3

    Match the task verb the prompt used

    Task verbs are instructions, not decoration, and three of them set the writing bar you have to clear. Identify or State means give the information without elaboration. Describe means say what the thing is or what it means in the context of this study. Explain means supply the because — the causal or logical link between two things. An Explain row scored against a naming sentence gets nothing, no matter how correct the name is. AAQ Part F is an Explain row, even though its prompt ends with the words supports or refutes. Picking a side there is not enough; you have to say why. A fourth verb, Propose, governs the EBQ claim in Part A. When you see Explain, type the word because or therefore into the sentence. It forces the link onto the page instead of leaving it in your head.

  4. 4

    Keep Identify answers short, but know why

    Correct extra material is not penalized. A reader does not dock you for a true add-on sentence, and credited sample responses on the research-method row often carry an elaboration. What costs you is an inaccurate statement sitting next to the correct one in the same part. So keep Identify answers short for two real reasons. It saves minutes, and every extra clause is one more chance to write something wrong. AAQ Part A, the research-method row, is the clean example. Naming the method is the whole task, so one sentence does it. Part D is not that row, even though it opens with Identify. Its prompt is one short line asking for an ethical guideline the researchers followed. The second half hides in the rubric, not in the prompt. To earn that 1 point you have to name the guideline and describe one way the researchers applied it. A sentence that only says the researchers obtained informed consent scores zero. So add the describing sentence every time. Naming a guideline the passage never shows happening also scores zero.

  5. 5

    Label every part and answer in order

    On the AAQ, type each part label, A through F, at the start of its own short paragraph. On the EBQ, use A, B(i), B(ii), C(i), C(ii). Do not mix the two schemes. Readers score row by row against the rubric; they are not searching your prose for a buried correct idea, and anything they cannot locate does not exist. Never fuse two parts into one paragraph, even when the answers overlap. Never answer out of order. One part, one paragraph, two to four sentences. If an AAQ paragraph runs past five sentences, you are padding.

  6. 6

    Use the term-plus-tether pattern in every scored sentence

    Each sentence that earns a point has two halves. The first half names the course concept exactly. The second half tethers it to a specific detail from this study — a number, a procedure, a group, a measurement. A definition floating on its own is the most common way strong students lose points, because the rubric asks for application, not recall. Compare a weak answer, that a correlational study measures the relationship between two variables, with a scoring one: this is a correlational study because the researchers measured both screen time and sleep duration as they already existed and manipulated neither. Same concept, one tether, one point.

  7. 7

    Build the EBQ around one claim and two sources

    Part A wants a claim you can defend. Write it as one sentence a reasonable person could argue against. A sentence nobody would dispute is not a claim, and restating the prompt is not a claim. Then pick the two sources that back it best. Part B(i) and Part C(i) have to draw on different sources from each other. In each, give that source's specific finding and cite it. You do not have to quote every decimal, but any number or finding you do report must match the source. The source is printed in front of you, so check it. Paraphrasing the direction of a finding is safe; misstating a value is not. Inaccurate evidence can cost the evidence point, and it can also cost the explanation row built on it. Put the evidence and its citation in a sentence of their own, then start the explanation in a fresh sentence. The reader can then see at a glance which sentence is doing the evidence work.

    In B(ii) and C(ii), explain how that evidence supports your claim, then name a psychological concept and apply it. Those are the two points: the link earns one, the applied concept earns the second. Read the evidence accurately, because an explanation built on a misread source can lose its credit however good the rest is. The concept has to be one you apply, not one you copy. Restating a source's own wording is not applying a concept. The perspective, theory, concept, or research finding you use must come from the AP Psychology course framework, so do not reach outside the course for it.

    Two more rules people miss. C(ii) has to apply a different concept than B(ii); recycle one and you score 1 instead of 2. And the C(i) evidence must come from a different source than the B(i) evidence. C(ii) explains that C(i) evidence, and the only new rule it adds is the different concept. If your C(ii) explanation also refers back to the B(i) source, that alone does not cost the row.

    Every row is scored on its own. So if you doubt your citation at B(i) or C(i), write the explanation part that follows it anyway. A lost evidence point does not sink the 2-point row after it.

  8. 8

    Audit for three minutes at the end

    Count your labeled answers against the number of prompts; a missing label is usually a missing answer. Confirm every EBQ evidence sentence carries a source number. Confirm B(ii) and C(ii) name two different psychological concepts. Then hunt three things. Causal language attached to a correlational study — rewrite as is associated with. Sentences that contradict something you wrote earlier — delete one of them. And hedges like arguably or this could possibly suggest. A committed wrong answer costs one row. A hedge that answers both ways costs the same row and buys nothing. A mixed verdict on generalizability is different, because that prompt invites one, and it scores when you back each half with evidence.

Worked through

Both stimuli below are invented for this guide. The studies are not real and every number in them is made up.

Here is a practice AAQ in the exam's format — one stimulus, six parts labeled A through F — with the scoring logic exposed. The part wordings below are our own paraphrases of the task types, not the exam's prompt lines.

The stimulus: Researchers tested whether brief aerobic activity improves memory in teenagers. They recruited 120 students aged 14 to 17 who volunteered in response to posters at two high schools in the same suburban district. Each student was randomly assigned either to walk briskly on a treadmill for 15 minutes or to sit quietly for 15 minutes. Immediately afterward, every student studied a list of 30 unrelated nouns for two minutes, then wrote down as many as they could recall in 90 seconds. The exercise group recalled a mean of 14.2 words; the seated group recalled a mean of 11.6 words. The difference was statistically significant at p < .01. Before taking part, students and their guardians signed forms describing the treadmill task and the recall test and stating that a student could stop at any point.

A. Name the research method this study used. Response: This was an experiment, because the researchers manipulated an independent variable — 15 minutes of treadmill walking versus 15 minutes of sitting — and randomly assigned students to those two conditions. Why it scores: Identify only needs the name. The added clause is correct, so it is neutral, and it makes clear you meant this study. Keep it to one sentence.

B. Give the operational definition of memory in this study. Response: Memory was operationally defined as the number of words from the 30-noun list that a student wrote down correctly during the 90-second recall period. Why it scores: an operational definition is the measurement procedure with its units, not the concept. Writing that memory is the retention of information over time earns nothing here.

C. Describe what the mean number of words recalled indicates about the two groups. Response: The means indicate that the exercise group recalled more words than the seated group, 14.2 words on average versus 11.6, a gap of 2.6 words. On this recall task, the students who walked for 15 minutes performed better than the students who sat. Why it scores: this row wants you to say what the statistic shows in context, with direction — which group scored higher, on what measure. Naming the statistic and stopping is not a description. Note the separate skill. If a prompt asks you to interpret the p value itself, translate it into probability language. Here, p < .01 means a difference this large or larger would turn up by chance less than 1% of the time if exercise had no real effect. Never treat significant as a synonym for large or important.

D. Name one ethical guideline the researchers followed. Response: The researchers obtained informed consent from the participants. Students and their guardians signed forms describing the treadmill task and the recall test, and stating that a student could stop at any point. Why it scores: the prompt is one line, but the rubric behind it wants two halves for its single point. The first sentence names the guideline. The second describes what the researchers actually did that shows it. Write the first sentence alone and you score zero. What also scores zero is naming a guideline the passage never shows. An answer saying the researchers debriefed the students after the recall test fails, because nothing in this study describes a debriefing. Naming a guideline the researchers should have used, but did not, earns nothing either.

E. Explain how far these findings can be generalized. Response: These findings do not generalize well to teenagers in general, because the participants were self-selected volunteers from just two high schools in one suburban district. Teens who choose to sign up for a study may differ from teens who do not, and students from one suburb may not represent urban or rural teenagers. Why it scores: the row needs a claim about generalizing to some population plus evidence taken from the participants themselves — who they were, where they came from, how they were recruited. Notice what the answer leads with: how the students were recruited and where they came from. Lead with sampling and participant characteristics rather than sample size, because noting that only 120 students took part is a weaker basis and may not be credited on its own. A mixed verdict is allowed. You can argue the findings partly generalize, as long as your evidence covers both sides. Keep it to one or two sentences and move on.

F. Use at least one finding from the study to explain whether it supports or refutes the claim that physiological arousal improves memory encoding. Response: The finding supports the claim, because the group that walked for 15 minutes recalled a mean of 14.2 words while the seated group recalled 11.6. Aerobic activity raises heart rate and physiological arousal. The Yerkes-Dodson law describes an inverted-U link between arousal and performance: moderate arousal improves performance on tasks like this recall test, and very high arousal hurts it. A brisk 15-minute walk sits in that helpful middle range, so the walkers were better prepared to encode the list during the two-minute study period. Why it scores: this row is worth 2, and it is an Explain row. Full marks need a specific result plus an explanation, with the result read accurately. You drop to 1 if you cite 14.2 versus 11.6 and never explain. You drop to 1 if you explain arousal and encoding and never cite a result. You also drop to 1 if you supply both but misread the data — calling a 2.6-word gap proof that exercise doubles memory, for instance.

Now the EBQ, worth the same 7 points across 45 minutes. These three sources are invented too, numbers included.

Here are the three sources, boiled down. Source 1 randomly assigned 240 ninth graders at three schools to either a two-session online growth mindset program or a study-skills control. By June the mindset group had attempted a mean of 8.4 optional challenge problems, against 6.4 for the control, a statistically significant gap. Source 2 surveyed 1,100 high school students; those who agreed that ability can be developed reported more homework time after a bad grade, r = .34. Source 3 is a meta-analysis of 39 mindset interventions, with an average effect of d = 0.10 overall that rises to d = 0.24 for students with low prior achievement. The question: should schools run growth mindset programs?

A. Make a specific claim you can defend. Response: Schools should run growth mindset programs, because they produce real gains in student effort, and the gains are largest for students who are already struggling. Why it scores: it takes one side, a reasonable person could argue the other, and it does not repeat the question back. One sentence is enough.

B(i). Back your claim with evidence from one source. Response: A meta-analysis of 39 mindset studies found an average effect of d = 0.10 across all students, rising to d = 0.24 — more than twice as large — among students with low prior achievement (Source 3). Why it scores: a specific finding, cited. The evidence and its citation sit in one sentence by themselves, and the explanation starts in a new sentence, so the reader can see exactly which sentence does the evidence work.

B(ii). Explain how that evidence supports your claim, using a psychological concept. Response: The overall effect is real but small, while the effect for low-achieving students is more than twice as large, so the program pays off most for the students schools most need to reach. Bandura's concept of self-efficacy explains why. Students with a history of failure doubt their own capability to do the task, so they put in little effort. A credible message that ability grows lifts that capability belief, which is exactly the one they are missing. Students who already expect to succeed have less room to move. Why it scores: 2 points. The first sentence links the evidence to the claim, which is one point. Self-efficacy is a psychological concept, named and applied to these particular students, which is the second. Stop after the first sentence and you earn 1. Misread d = 0.10 and d = 0.24 and you can lose the explanation credit, because the explanation has to rest on an accurate reading of the evidence. Watch what would not earn the concept point here. Effect size and meta-analysis are statistics vocabulary, not psychological concepts. Growth mindset is a real course concept, but repeating the phrase as Source 1 already words it is quoting, not applying. Self-efficacy is the safer move, because you supply the mechanism yourself.

C(i). Back your claim with evidence from a second, different source. Response: Ninth graders who completed a two-session mindset program attempted a mean of 8.4 optional challenge problems by June, compared with 6.4 in the control group (Source 1). Why it scores: it uses a different source from B(i), which C(i) requires, and it reports a specific finding with the direction clear.

C(ii). Explain, applying a different concept than the one in B(ii). Response: Attempting harder optional problems is the behavior you would predict if the program changed how students explain their own failures. Attribution theory says a student who blames a bad grade on a fixed internal trait quits, while a student who blames an unstable, controllable cause such as strategy keeps working. Those two extra problems per student are that shift in attribution showing up as behavior. That is why even a two-session program given to a whole grade is worth running. Why it scores: 2 points again. Attribution theory is genuinely different from self-efficacy, so the concept point is live. Reuse self-efficacy here and the row caps at 1.

Common mistakes

  • Defining the term instead of applying it. The rubric asks what this study did, so a textbook definition with no number, group, or procedure from the passage attached to it earns nothing.

  • Mislabeling the parts. The AAQ is A through F with no roman numerals; the EBQ is A, B(i), B(ii), C(i), C(ii). Writing A(i) on an AAQ, writing a bare B or C on the EBQ with no sub-label, or calling an EBQ source Source A instead of Source 1, sends a reader hunting for answers they will not credit.

  • Treating Part D as a naming row, or naming a guideline the study never shows. The prompt is one short line asking for an ethical guideline the researchers followed, but the rubric behind it wants the name plus a description of what they did. The most common Part D zero is naming a real guideline from the study and stopping. Writing that they debriefed the students scores zero when the passage describes no debriefing. So does naming a guideline the researchers should have followed but did not.

  • Using causal language about a correlational study. If nobody manipulated a variable, the finding shows an association or a relationship. Words like caused, led to, made, and increased hand the reader a reason to withhold the point.

  • Confusing random sampling with random assignment, then using the wrong one as evidence. Random assignment makes groups equivalent and licenses a causal claim. Random sampling makes the sample representative and licenses generalization. Build a generalizability answer on how participants were sampled and who they were; sample size alone is a weaker basis and may not be credited.

  • Naming a statistic instead of describing what it shows. On the statistic-interpretation row, say which group scored higher on what measure, in the context of this study. Save probability language for a prompt that actually asks about a p value.

  • On the EBQ, giving evidence without citing its source number, or citing a source and stopping. Every evidence sentence needs a citation, and each explanation part needs both the link to your claim and an applied psychological concept, theory, perspective, or research finding.

  • Reusing in C(ii) the concept you already applied in B(ii), or parroting a source's own wording instead of applying an idea. Reuse caps that row at 1 point instead of 2. The concept itself must be a perspective, theory, concept, or research finding from the AP Psychology course framework, applied to the evidence in your own words.

  • Writing in fragments, bullets, or unlabeled blocks. Complete sentences are required, and an answer a reader cannot locate is an answer that does not exist.

Checklist

  • Can you name the research method and justify it from the study's procedure in a single sentence?
  • Can you write an operational definition that contains the actual measurement, its units, and its time window, with no textbook definition in it?
  • Can you describe what a difference in two group means indicates — which group was higher, on what measure, by how much — in one sentence, and separately say what a p value like p < .01 means in probability language?
  • Can you name four ethical guidelines from memory, tell which of them a given passage actually shows the researchers applying, and then describe in one sentence how they applied it in that study?
  • Can you earn the generalizability point in one or two sentences by naming a population and citing who the participants were and how they were recruited, without leaning on random assignment or on sample size alone?
  • Can you write an EBQ claim that takes one side, could be argued against, and does not restate the prompt?
  • Can you write an evidence sentence that ends with its citation — (Source 2) or an According to Source 2 lead-in — and states that source's specific finding accurately?
  • Can you write two explanation paragraphs, B(ii) and C(ii), that each link evidence to your claim and each apply a different concept from the AP Psychology course framework that you supply yourself rather than lift from a source?
  • On AAQ Part F, can you produce both halves — a specific result from the study and an explanation tying it to the concept named in the prompt — and read the result accurately, to reach 2 points?
  • Can you finish the AAQ in 25 minutes with all six parts labeled A through F, in order, and answered in complete sentences?

More AP Psychology technique