Skip to content

Free-response practice

The Article Analysis Question (AAQ)

7 points2 original prompts4 scored examples

The Article Analysis Question is Question 1 of Section II. You get a short summary of one research study, written like the abstract of a journal article. It tells you who the participants were, how they were recruited, what the researchers did, what they measured, and what they found. Sometimes the results come as a small table or graph. Then you answer six lettered parts, A through F, for a total of 7 points.

On this page · 6 sections

Section II gives you 70 minutes for two questions and counts for 33.3 percent of your AP score. The AAQ and the Evidence-Based Question are worth 7 points each, which makes the AAQ about one-sixth of your score. Plan on roughly 25 minutes for it, since the EBQ needs the remaining time for its three sources.

The six parts follow a fixed pattern, so learn it now and nothing on the page will surprise you. - Part A: identify the research method. - Part B: state the operational definition of a variable. - Part C: describe what a reported statistic means in this study. - Part D: identify an ethical guideline the researchers followed. - Part E: explain how far the findings can be generalized, using details of the sample. - Part F: explain how a finding supports or refutes a hypothesis or claim.

Parts A through E are worth 1 point each, and Part F is worth 2.

Readers score each part independently, with no thesis, no essay structure and no credit for style. A part either does the named task or it does not, and missing Part A does not cost you Part B. Answer every part, even when you are uncertain.

Two habits keep points safe. First, label your answers A through F and answer them in order. Second, write full sentences that tie each answer to the study. "Experiment" alone is a guess. "This was an experiment because participants were randomly assigned to cycle or read" is an answer. Everything you cite must come from the summary itself, because outside knowledge of the actual study is neither needed nor rewarded.

The rubric, in plain terms

7 points, earned line by line. Readers score what is on the page, nothing else.

Part A: research method1 pt

Names the method the study actually used, such as experiment, correlational study, case study, naturalistic observation or survey. The safest answers add the feature that proves it: for an experiment, the manipulated variable and random assignment; for a correlational study, the fact that both variables were measured and nothing was assigned. Naming two methods, one right and one wrong, does not earn the point.

Part B: operational definition1 pt

States exactly how the named variable was measured or manipulated in this study. That means the procedure and the unit, such as seconds before the hand came out of the water, or yellow lights driven through out of 20. Repeating the concept in other words ("how much pain someone can handle") or naming the variable type ("it was the dependent variable") does not earn the point.

Part C: meaning of the statistic1 pt

Says accurately what the reported number tells you about these results. For a correlation, give both direction and strength, and say what the direction means for the two variables. For an effect size, say how large the difference between groups is. For a significance value, say how unlikely a difference this big would be if chance alone were at work. Mixing up the statistics (calling an effect size "significance") or reading the direction backward loses the point.

Part D: ethical guideline1 pt

Names a recognized research-ethics guideline and connects it to something the summary says the researchers did. Examples are informed consent (which includes the right to withdraw), assent for minors, confidentiality, protection from harm and debriefing. Naming a guideline the study does not show, or describing a procedure without naming any guideline, does not earn the point.

Part E: generalizability1 pt

Takes a position on how far the results extend and backs it with a specific characteristic of the participants or how they were chosen. Age, nationality, school, student status, health screening and recruitment method all qualify as evidence. The answer must say which wider group the results may or may not apply to, and why that characteristic matters. A bare complaint that the sample was small names no group and does not do this job.

Part F: finding supports or refutes the claim2 pts

First point: cites a specific, accurately reported result from the study, such as group means, a correlation or a comparison between subgroups. Second point: explains how that result supports or refutes the stated hypothesis or claim, in a way that is accurate psychology. An explanation that draws a conclusion the design cannot support, such as causation from correlational data, or that leans on a wrong mechanism, does not earn the second point.

Practice prompts

Prompt 1

Does a bike ride raise pain tolerance?

Researchers at a university in Canada tested whether aerobic exercise increases pain tolerance. They recruited 64 undergraduates (40 women and 24 men, ages 18 to 23) from an introductory psychology course. Students received course credit for taking part. Anyone with a heart condition, high blood pressure or a chronic pain condition was screened out before the study began. Every participant signed a consent form that described the cold-water task. Each was told they could stop at any time without losing credit.

Participants were randomly assigned to one of two conditions, 32 per group. The exercise group rode a stationary bike for 20 minutes at moderate intensity, with heart rate kept between 60 and 70 percent of each person's estimated maximum. The control group sat and read a magazine for 20 minutes.

Immediately afterward, each participant placed one hand in water held at 4°C (39°F). Pain tolerance was recorded as the number of seconds before the participant chose to remove the hand. For safety, the task ended automatically at 180 seconds.

Results: The exercise group kept their hands in the water for an average of 92 seconds (SD = 31). The control group averaged 61 seconds (SD = 28). The difference was statistically significant (p < .01). The effect size was d = 1.05. After the session, every participant was told the full purpose of the study and given a chance to ask questions.

  1. (A)Identify the research method used in this study. [1 pt]
  2. (B)State the operational definition of pain tolerance in this study. [1 pt]
  3. (C)Describe what the effect size of d = 1.05 indicates about the results. [1 pt]
  4. (D)Identify one ethical guideline the researchers applied, and show where it appears in the study. [1 pt]
  5. (E)Explain the extent to which the findings of this study can be generalized, using specific and relevant evidence from the study. [1 pt]
  6. (F)Explain how at least one of the research findings supports or refutes the researchers' hypothesis that aerobic exercise increases pain tolerance. [2 pts]
Prompt 2

Short sleep and risky choices in teenagers

Researchers in Ohio studied whether adolescents who sleep less make riskier decisions. Their reasoning was that the prefrontal cortex, which supports impulse control, is still developing in the teen years. They expected short sleep to weaken it further.

The researchers mailed letters to families at three public high schools in one Ohio county. They enrolled 212 students, ages 14 to 17. A parent of each student signed a consent form, and each student also agreed to take part. Students' names were replaced with ID numbers on all data files.

For seven school nights in a row, each student wore a wrist device that recorded how long they slept. Average nightly sleep ranged from 4.9 to 8.9 hours, with a mean of 6.8 hours. On the eighth day, each student played a 10-minute computer driving game. The game had 20 intersections where the light turned yellow as the car approached. Risk-taking was scored as the number of intersections, out of 20, where the student kept driving instead of stopping.

Results: The correlation between average nightly sleep and the risk-taking score was r = −0.41 (p < .001). Students who averaged fewer than 7 hours of sleep drove through an average of 9.1 yellow lights. Students who averaged 7 hours or more drove through an average of 5.6.

  1. (A)Identify the research method used in this study. [1 pt]
  2. (B)State the operational definition of risk-taking in this study. [1 pt]
  3. (C)Describe what the correlation coefficient of r = −0.41 indicates about the relationship between sleep and risk-taking. [1 pt]
  4. (D)Identify one ethical guideline the researchers applied, and show where it appears in the study. [1 pt]
  5. (E)Explain the extent to which the findings of this study can be generalized, using specific and relevant evidence from the study. [1 pt]
  6. (F)Explain how at least one of the research findings supports or refutes the researchers' hypothesis that adolescents who get less sleep make riskier decisions. [2 pts]

Scored examples

Two responses per prompt — one near the top of the rubric, one in the middle — with every point accounted for. The mid response earns its score through the mistakes students actually make.

Does a bike ride raise pain tolerance? · high-scoring response7/7

A. This study was an experiment. The researchers manipulated whether people exercised, and they randomly assigned participants to the bike group or the reading group.

B. Pain tolerance was defined as the number of seconds a participant kept one hand in 4°C water before choosing to pull it out, with a maximum of 180 seconds.

C. An effect size of d = 1.05 means the exercise group's average was about one standard deviation higher than the control group's. That is a large difference, bigger than the usual 0.8 cutoff for a large effect. So exercise did not just make a statistically real difference. It made a big one.

D. The researchers used informed consent. Before starting, every participant signed a form that explained the cold-water task and told them they could stop at any time.

E. The findings may generalize to healthy college students around 18 to 23, because that is who was tested. They may not generalize to older adults or to people with chronic pain or heart conditions. Those people were screened out, and their pain systems or exercise ability could be different.

F. The exercise group kept their hands in the cold water for 92 seconds on average, while the reading group lasted only 61 seconds. This supports the hypothesis. Because participants were randomly assigned, the groups should have started out similar, so the extra 31 seconds can be traced to the 20 minutes of cycling. Exercise may raise endorphins, which bind to opioid receptors and dampen pain signals. This study did not measure endorphins, though, so it shows that exercise increased tolerance, not exactly why.

How it scored

1/1A — Earned. Names the experiment and points to both features that make it one: a manipulated variable and random assignment.
1/1B — Earned. Gives the procedure and the unit, seconds in 4°C water, so a reader could repeat the measurement.
1/1C — Earned. Correctly reads d as the size of the gap in standard-deviation units and calls it large. It also keeps effect size separate from significance, which many students blur.
1/1D — Earned. Names informed consent and ties it to the signed form described in the study.
1/1E — Earned. Uses two real features of the sample, the age range and the health screening. It says which groups the results may not reach and why.
2/2F — Both earned. The first point comes from the accurate means, 92 versus 61 seconds. The second comes from explaining why random assignment lets the difference be credited to exercise. The endorphin note is accurate, and the student correctly says the study did not test it.
Does a bike ride raise pain tolerance? · mid-scoring response3/7

A. Experiment, because there was an experimental group that rode a bike and a control group that read.

B. Pain tolerance is how much pain a person can handle before it gets too uncomfortable.

C. The effect size of 1.05 means the results were statistically significant, so there is less than a 5% chance they happened by accident.

D. Informed consent. The participants signed a consent form before they did the cold water task.

E. The results are not very generalizable because only 64 people were in the study. That is too small to apply to everyone.

F. The bike group kept their hands in the water for 92 seconds and the reading group only kept them in for 61 seconds. This supports the hypothesis because exercise activates the parasympathetic nervous system, which calms the body and slows the heart rate, so the person does not feel the pain as much.

How it scored

1/1A — Earned. Experiment is correct, and the response points to the two conditions. Mentioning random assignment would have made this airtight.
0/1B — Not earned. This restates what pain tolerance means in general. It never says how this study measured it: seconds in 4°C water, capped at 180.
0/1C — Not earned. The response treats effect size as a significance test. The p-value speaks to chance. The effect size says how big the gap is: about one standard deviation, which is large.
1/1D — Earned. Names a guideline and links it to the signed form in the study.
0/1E — Not earned. Sample size alone does not name any group the results might fail to reach. The study gives better evidence: all participants were 18 to 23, all were students in one course, and anyone with heart or pain conditions was excluded.
1/2F — First point earned for the accurate means. Second point lost because the mechanism is wrong. Moderate cycling activates the sympathetic system, which raises heart rate. Slowing the heart is a parasympathetic, rest-and-digest effect. The response also never uses the design: with random assignment, the extra seconds can be credited to the exercise.
Short sleep and risky choices in teenagers · high-scoring response6/7

A. This was a correlational study. The researchers measured two things that already varied, sleep and risk-taking, and looked at how they were related. They did not assign anyone to sleep more or less.

B. Risk-taking was how willing the student was to take chances and do dangerous things.

C. An r of −0.41 is a moderate negative correlation. As average sleep went down, the number of yellow lights driven through tended to go up. It is not a perfect relationship, so sleep only partly predicts the score, and it does not show that one causes the other.

D. Confidentiality. The researchers replaced each student's name with an ID number on the data files, so no one's results could be traced back to them.

E. The results may generalize to high school students ages 14 to 17 in similar American public schools, since the sample came from three of them. They may not generalize to adults, whose prefrontal cortex is more mature, or to teens in other countries or school systems. All students came from one Ohio county, and families who answered a mailed letter may differ from families who did not.

F. Students who slept under 7 hours ran an average of 9.1 yellow lights, compared with 5.6 for students who slept 7 hours or more. This supports the hypothesis, because the teens who slept less did make more risky choices, just as predicted. Since this is correlational, though, it only shows that less sleep goes along with more risk. A third variable, like a more impulsive personality, could cause both staying up late and running lights.

How it scored

1/1A — Earned. Names the correlational method and gives the reason: both variables were measured, nothing was assigned.
0/1B — Not earned. This restates the concept of risk-taking in everyday words. It never says how the study measured it: the number of intersections, out of 20, where the student kept driving on a yellow light.
1/1C — Earned. Gives direction and strength, translates the direction into the two variables, and correctly avoids a causal reading.
1/1D — Earned. Names confidentiality and ties it to the ID numbers in the study.
1/1E — Earned. Uses the age range, the single county and the mailed-letter recruitment. It explains which groups the results may not reach, with a reason tied to development.
2/2F — Both earned. The subgroup means are cited accurately. The explanation says the pattern matches the prediction and stays within what correlational data can show, naming a plausible third variable.
Short sleep and risky choices in teenagers · mid-scoring response4/7

A. Correlational study, because they measured sleep and risk-taking to see if they were related.

B. Risk-taking was the number of yellow lights out of 20 that the student drove through in the computer game.

C. r = −0.41 means there is a weak negative correlation. Since it is negative, when students got less sleep they also took fewer risks.

D. Confidentiality, because the students' names were changed to ID numbers.

E. The results can be generalized to all teenagers because 212 students is a big sample and they came from three different schools.

F. The correlation was r = −0.41, which supports the hypothesis. It proves that losing sleep causes teenagers to take more risks because the prefrontal cortex cannot control impulses when you are tired.

How it scored

1/1A — Earned. Correct method with a reason tied to the study.
1/1B — Earned. Gives the exact measure and its scale, out of 20 intersections.
0/1C — Not earned. The direction is read backward. A negative correlation means the variables move in opposite directions: less sleep went with more yellow lights, not fewer. Also, −0.41 is usually called moderate, not weak.
1/1D — Earned. Names a guideline and connects it to the ID numbers the study describes.
0/1E — Not earned. Claiming the results reach all teenagers ignores what the summary says about the sample: ages 14 to 17, one Ohio county, families who answered a mailed letter. Size and three schools do not make a sample representative of every teenager.
1/2F — First point earned for citing the correlation accurately. Second point lost because the explanation claims causation. Nothing was manipulated, so the study cannot show that sleep loss causes risk-taking. The prefrontal account may be right, but this design cannot prove it.

Where points die

Writing a definition of the concept in Part B instead of the study's measure, such as "pain tolerance is how much pain you can take." An operational definition is a measurement procedure, so copy the measurement from the summary: seconds before the hand came out of 4°C water, or yellow lights driven through out of 20. If someone could not repeat the study from your sentence, it is not operational.

Treating an effect size as a significance test, or reading a negative correlation as "weak" or as both variables falling together. Keep the three statistics apart. A p-value says how unlikely the result would be if chance alone were at work. An effect size like d = 1.05 says how big the gap is, here about one standard deviation, which is large. A correlation's sign gives direction and its size gives strength. So −0.41 means moderate, with the variables moving in opposite directions.

Saying a correlational study "proves" that one variable causes the other in Part F. Only an experiment, with a manipulated independent variable and random assignment, supports a causal claim. If nothing was manipulated, say the finding is consistent with the hypothesis and name a possible third variable. Overclaiming costs the explanation point even when the finding you cite is correct.

Answering Part E with "the sample was too small" or "the sample was big, so it applies to everyone." Size alone won't earn the point, because generalizing depends mainly on who the participants were and how they were chosen. Pull a characteristic from the summary: age range, one county, college course credit, health screening, recruitment by mail. Then name a group the results may not reach and say why that difference could matter.

Using the wrong biology to explain a result, such as crediting the parasympathetic system for exercise effects, or saying endorphins work through serotonin. Exercise and stress drive the sympathetic division: faster heart rate, more alertness, energy released for action. Rest-and-digest effects belong to the parasympathetic division. Endorphins reduce pain by binding to opioid receptors, the same ones morphine acts on. A wrong mechanism in Part F makes the explanation inaccurate, and the second point goes with it.

Naming an ethical guideline in Part D that the study never shows, such as debriefing when the summary says nothing about it, or describing a step without naming the guideline. Find the sentence in the summary first, then name it: a signed form is informed consent, a parent's signature plus the teen's agreement is consent and assent, ID numbers are confidentiality, a stop-anytime rule is part of informed consent (the right to withdraw), a 180-second cap is protection from harm. One named guideline tied to one detail is all the point needs.

Score your own response

0 / 7

Be the harshest reader in the room: award a point only when your response does what the line says, on the page — not what you meant to write.