The revised AP Statistics Exam has three hours of testing time: 90 minutes for 42 multiple-choice questions and 90 minutes for four free-response questions. Administrative instructions, sign-in, transitions, a scheduled break, and any approved accommodations can make the student’s total time at the testing site longer than three hours.
How long is the AP Statistics Exam?
Exactly 180 minutes of section testing time. Section I is 90 minutes and Section II is 90 minutes. Both sections contribute 50% of the score. The testing total should not be confused with the student’s reporting-to-dismissal time.
The revised counts apply to the May 2027 fully digital exam. Older resources may say 40 MCQs and six FRQs; those describe the previous structure. They still show useful statistics problems, but they do not provide the current pacing denominator.
AP Statistics section timeline
| Part | Official testing time | Questions | Score weight | Pacing implication |
|---|---|---|---|---|
| Section I: Multiple Choice | 90 minutes | 42 | 50% | Average 2.14 minutes per item; use cumulative checkpoints rather than a rigid item timer |
| Section II: Free Response | 90 minutes | 4 questions at 10 points each | 50% | Average 22.5 minutes, but question roles and parts differ |
| Total | 180 minutes | 46 questions/items across both sections | 100% | Balance recognition, computation, written reasoning, and interpretation |
Testing time is not the same as total room time
The three-hour figure counts the two timed sections. Students also follow proctor instructions, sign in to the testing workflow, receive materials, transition between sections, take any scheduled break, and complete submission procedures. The school controls reporting and dismissal communication, so no universal article should promise an exact departure time.
Transportation plans should begin with the coordinator’s reporting time and end with the school’s estimated dismissal window. Adding 180 minutes to a published morning or afternoon session can be inaccurate because the session label is not the arrival time and administration does not begin the instant a student enters the room.
Device or network issues are handled through the authorized testing process and may affect a student’s room experience without changing the official section allocations. During practice, students should rehearse setup separately so their statistical pacing data are not contaminated by avoidable account or hardware problems.
The mathematics of MCQ pacing
Dividing 90 minutes by 42 questions gives approximately 2.1429 minutes per question, or about 2 minutes 9 seconds. That is an average over the section, not a command to stop at exactly that time on every item.
| Checkpoint | Target elapsed | Target remaining | Meaning |
|---|---|---|---|
| Question 7 | 15 minutes | 75 minutes | One-sixth of the items complete |
| Question 14 | 30 minutes | 60 minutes | One-third complete |
| Question 21 | 45 minutes | 45 minutes | Half complete |
| Question 28 | 60 minutes | 30 minutes | Two-thirds complete |
| Question 35 | 75 minutes | 15 minutes | Five-sixths complete |
| Question 42 | 90 minutes | 0 minutes | Section complete |
A student who is two questions behind at the first checkpoint does not need to rush blindly. The repair is to reduce unproductive rereading and move past a true blocker. A student who is ahead should preserve reasoning quality and use the cushion on flagged calculations or interpretation choices.
A 42-question pacing map
The topics are a balanced rehearsal sequence, not a prediction of operational question order. The elapsed and remaining columns are exact proportional targets based on 90/42.
| Question | Target elapsed | Time remaining | Rehearsal focus | Decision |
|---|---|---|---|---|
| 1 | 2.1 min | 87.9 min | investigative question | First identify the population, variables, and answerable comparison before reading options. If the remaining work is only calculator entry or one comparison, finish it; otherwise mark it, choose the best supported option, and answer and move. |
| 2 | 4.3 min | 85.7 min | categorical display | First compare proportions with a common denominator rather than raw counts from unequal groups. If the remaining work is only calculator entry or one comparison, finish it; otherwise mark it, choose the best supported option, and answer and move. |
| 3 | 6.4 min | 83.6 min | quantitative distribution | First scan shape, center, spread, and unusual features before calculating. If the remaining work is only calculator entry or one comparison, finish it; otherwise mark it, choose the best supported option, and answer and move. |
| 4 | 8.6 min | 81.4 min | summary statistic | First choose resistant or nonresistant summaries to match shape and outliers. If the remaining work is only calculator entry or one comparison, finish it; otherwise mark it, choose the best supported option, and answer and move. |
| 5 | 10.7 min | 79.3 min | linear transformation | First separate what happens to location from what happens to spread. If the remaining work is only calculator entry or one comparison, finish it; otherwise mark it, choose the best supported option, and answer and move. |
| 6 | 12.9 min | 77.1 min | z-score | First standardize with the correct mean and SD, then interpret relative position. If the remaining work is only calculator entry or one comparison, finish it; otherwise mark it, choose the best supported option, and answer and move. |
| 7 | 15.0 min | 75.0 min | sampling method | First name the random mechanism and decide which population the frame represents. If the remaining work is only calculator entry or one comparison, finish it; otherwise mark it, choose the best supported option, and brief checkpoint: compare actual time with target. |
| 8 | 17.1 min | 72.9 min | survey bias | First identify a systematic design flaw that sample size cannot repair. If the remaining work is only calculator entry or one comparison, finish it; otherwise mark it, choose the best supported option, and answer and move. |
| 9 | 19.3 min | 70.7 min | experimental design | First separate random assignment for causation from random sampling for generalization. If the remaining work is only calculator entry or one comparison, finish it; otherwise mark it, choose the best supported option, and answer and move. |
| 10 | 21.4 min | 68.6 min | blocking or pairing | First use the matching structure and do not analyze linked observations as independent. If the remaining work is only calculator entry or one comparison, finish it; otherwise mark it, choose the best supported option, and answer and move. |
| 11 | 23.6 min | 66.4 min | simulation | First check the label assignment and recorded statistic before counting repetitions. If the remaining work is only calculator entry or one comparison, finish it; otherwise mark it, choose the best supported option, and answer and move. |
| 12 | 25.7 min | 64.3 min | addition rule | First draw or define overlapping events so the intersection is not counted twice. If the remaining work is only calculator entry or one comparison, finish it; otherwise mark it, choose the best supported option, and answer and move. |
| 13 | 27.9 min | 62.1 min | conditional probability | First restrict the denominator to the group after the conditioning bar. If the remaining work is only calculator entry or one comparison, finish it; otherwise mark it, choose the best supported option, and answer and move. |
| 14 | 30.0 min | 60.0 min | independence | First verify with a conditional or multiplication rule rather than intuition. If the remaining work is only calculator entry or one comparison, finish it; otherwise mark it, choose the best supported option, and brief checkpoint: compare actual time with target. |
| 15 | 32.1 min | 57.9 min | expected value | First define net outcomes and compute the probability-weighted long-run average. If the remaining work is only calculator entry or one comparison, finish it; otherwise mark it, choose the best supported option, and answer and move. |
| 16 | 34.3 min | 55.7 min | binomial conditions | First verify fixed n, two outcomes, independence, and constant p. If the remaining work is only calculator entry or one comparison, finish it; otherwise mark it, choose the best supported option, and answer and move. |
| 17 | 36.4 min | 53.6 min | binomial calculation | First look for an efficient complement before entering a long probability sum. If the remaining work is only calculator entry or one comparison, finish it; otherwise mark it, choose the best supported option, and answer and move. |
| 18 | 38.6 min | 51.4 min | normal area | First sketch the requested region and confirm left, right, or middle probability. If the remaining work is only calculator entry or one comparison, finish it; otherwise mark it, choose the best supported option, and answer and move. |
| 19 | 40.7 min | 49.3 min | normal percentile | First find z first, then return to the original scale and units. If the remaining work is only calculator entry or one comparison, finish it; otherwise mark it, choose the best supported option, and answer and move. |
| 20 | 42.9 min | 47.1 min | sampling distribution of p-hat | First distinguish p, p-hat, and p0 while checking large counts. If the remaining work is only calculator entry or one comparison, finish it; otherwise mark it, choose the best supported option, and answer and move. |
| 21 | 45.0 min | 45.0 min | one-proportion interval | First use observed counts and interpret confidence as long-run method performance. If the remaining work is only calculator entry or one comparison, finish it; otherwise mark it, choose the best supported option, and brief checkpoint: compare actual time with target. |
| 22 | 47.1 min | 42.9 min | proportion sample size | First use a planning value, compute, and round upward. If the remaining work is only calculator entry or one comparison, finish it; otherwise mark it, choose the best supported option, and answer and move. |
| 23 | 49.3 min | 40.7 min | one-proportion test | First use p0 in the null standard error and the prespecified tail. If the remaining work is only calculator entry or one comparison, finish it; otherwise mark it, choose the best supported option, and answer and move. |
| 24 | 51.4 min | 38.6 min | two-proportion interval | First define p1-p2 and use separate sample estimates in the standard error. If the remaining work is only calculator entry or one comparison, finish it; otherwise mark it, choose the best supported option, and answer and move. |
| 25 | 53.6 min | 36.4 min | two-proportion test | First pool successes only because the null sets the population proportions equal. If the remaining work is only calculator entry or one comparison, finish it; otherwise mark it, choose the best supported option, and answer and move. |
| 26 | 55.7 min | 34.3 min | chi-square expected count | First multiply marginal totals and divide by the grand total. If the remaining work is only calculator entry or one comparison, finish it; otherwise mark it, choose the best supported option, and answer and move. |
| 27 | 57.9 min | 32.1 min | chi-square conclusion | First say association, not linear correlation or causation. If the remaining work is only calculator entry or one comparison, finish it; otherwise mark it, choose the best supported option, and answer and move. |
| 28 | 60.0 min | 30.0 min | sampling distribution of x-bar | First use sigma over square root n and justify the normal approximation. If the remaining work is only calculator entry or one comparison, finish it; otherwise mark it, choose the best supported option, and brief checkpoint: compare actual time with target. |
| 29 | 62.1 min | 27.9 min | t distribution | First recognize unknown sigma, use s, and track degrees of freedom. If the remaining work is only calculator entry or one comparison, finish it; otherwise mark it, choose the best supported option, and answer and move. |
| 30 | 64.3 min | 25.7 min | one-mean interval | First check randomness, independence, shape, and outliers before interpreting. If the remaining work is only calculator entry or one comparison, finish it; otherwise mark it, choose the best supported option, and answer and move. |
| 31 | 66.4 min | 23.6 min | one-mean test | First write hypotheses about mu and state evidence at the chosen alpha. If the remaining work is only calculator entry or one comparison, finish it; otherwise mark it, choose the best supported option, and answer and move. |
| 32 | 68.6 min | 21.4 min | paired t procedure | First form one difference per pair and check the difference distribution. If the remaining work is only calculator entry or one comparison, finish it; otherwise mark it, choose the best supported option, and answer and move. |
| 33 | 70.7 min | 19.3 min | two-mean t procedure | First keep independent groups separate and do not force equal variances. If the remaining work is only calculator entry or one comparison, finish it; otherwise mark it, choose the best supported option, and answer and move. |
| 34 | 72.9 min | 17.1 min | Type I and II errors | First translate each wrong decision into the context before choosing. If the remaining work is only calculator entry or one comparison, finish it; otherwise mark it, choose the best supported option, and answer and move. |
| 35 | 75.0 min | 15.0 min | power | First connect effect size, n, variability, and alpha without changing alpha after data. If the remaining work is only calculator entry or one comparison, finish it; otherwise mark it, choose the best supported option, and brief checkpoint: compare actual time with target. |
| 36 | 77.1 min | 12.9 min | scatterplot | First describe direction, form, strength, and unusual points before using r. If the remaining work is only calculator entry or one comparison, finish it; otherwise mark it, choose the best supported option, and answer and move. |
| 37 | 79.3 min | 10.7 min | correlation | First interpret linear strength without causal language or sensitivity blindness. If the remaining work is only calculator entry or one comparison, finish it; otherwise mark it, choose the best supported option, and answer and move. |
| 38 | 81.4 min | 8.6 min | regression slope | First state predicted response change per one explanatory unit with units. If the remaining work is only calculator entry or one comparison, finish it; otherwise mark it, choose the best supported option, and answer and move. |
| 39 | 83.6 min | 6.4 min | residual | First compute observed minus predicted and interpret the sign. If the remaining work is only calculator entry or one comparison, finish it; otherwise mark it, choose the best supported option, and answer and move. |
| 40 | 85.7 min | 4.3 min | r-squared | First name the response variation explained by the fitted linear relationship. If the remaining work is only calculator entry or one comparison, finish it; otherwise mark it, choose the best supported option, and answer and move. |
| 41 | 87.9 min | 2.1 min | influence | First distinguish high leverage, large residual, and change after refitting. If the remaining work is only calculator entry or one comparison, finish it; otherwise mark it, choose the best supported option, and answer and move. |
| 42 | 90.0 min | 0.0 min | mixed reasoning | First identify the parameter, design, method, and allowable conclusion before computing. If the remaining work is only calculator entry or one comparison, finish it; otherwise mark it, choose the best supported option, and brief checkpoint: compare actual time with target. |
A 42-question error debrief
Use this after a timed set. It converts each topic into a precise repair and prevents the empty label careless from replacing diagnosis.
| Question | Topic | First diagnosis | Repair | Proof on the next set |
|---|---|---|---|---|
| 1 | investigative question | If this item was slow or wrong, classify the first failed step as reading; do not label the whole result 'careless.' | Repair it by having the student underline the population, variable, comparison, and requested direction before looking at choices. Then solve a contrasting example in which a tempting alternative method would be wrong. | On an unseen investigative question item, the student should identify the population, variables, and answerable comparison before reading options. The answer must be reached inside the 2.14-minute average without omitting the controlling definition or condition. |
| 2 | categorical display | If this item was slow or wrong, classify the first failed step as method selection; do not label the whole result 'careless.' | Repair it by having the student write the parameter and data structure, then name why the competing procedure does not apply. Then solve a contrasting example in which a tempting alternative method would be wrong. | On an unseen categorical display item, the student should compare proportions with a common denominator rather than raw counts from unequal groups. The answer must be reached inside the 2.14-minute average without omitting the controlling definition or condition. |
| 3 | quantitative distribution | If this item was slow or wrong, classify the first failed step as condition; do not label the whole result 'careless.' | Repair it by having the student state the exact randomness, independence, count, or distribution check that controls the method. Then solve a contrasting example in which a tempting alternative method would be wrong. | On an unseen quantitative distribution item, the student should scan shape, center, spread, and unusual features before calculating. The answer must be reached inside the 2.14-minute average without omitting the controlling definition or condition. |
| 4 | summary statistic | If this item was slow or wrong, classify the first failed step as calculator entry; do not label the whole result 'careless.' | Repair it by having the student estimate magnitude first, write the intended function and parameters, then compare the display with the estimate. Then solve a contrasting example in which a tempting alternative method would be wrong. | On an unseen summary statistic item, the student should choose resistant or nonresistant summaries to match shape and outliers. The answer must be reached inside the 2.14-minute average without omitting the controlling definition or condition. |
| 5 | linear transformation | If this item was slow or wrong, classify the first failed step as algebra or arithmetic; do not label the whole result 'careless.' | Repair it by having the student keep extra digits, track parentheses and signs, and substitute the result back into the original expression. Then solve a contrasting example in which a tempting alternative method would be wrong. | On an unseen linear transformation item, the student should separate what happens to location from what happens to spread. The answer must be reached inside the 2.14-minute average without omitting the controlling definition or condition. |
| 6 | z-score | If this item was slow or wrong, classify the first failed step as interpretation; do not label the whole result 'careless.' | Repair it by having the student name the population, units, direction, uncertainty, and design limit in a complete contextual sentence. Then solve a contrasting example in which a tempting alternative method would be wrong. | On an unseen z-score item, the student should standardize with the correct mean and SD, then interpret relative position. The answer must be reached inside the 2.14-minute average without omitting the controlling definition or condition. |
| 7 | sampling method | If this item was slow or wrong, classify the first failed step as reading; do not label the whole result 'careless.' | Repair it by having the student underline the population, variable, comparison, and requested direction before looking at choices. Then solve a contrasting example in which a tempting alternative method would be wrong. | On an unseen sampling method item, the student should name the random mechanism and decide which population the frame represents. The answer must be reached inside the 2.14-minute average without omitting the controlling definition or condition. |
| 8 | survey bias | If this item was slow or wrong, classify the first failed step as method selection; do not label the whole result 'careless.' | Repair it by having the student write the parameter and data structure, then name why the competing procedure does not apply. Then solve a contrasting example in which a tempting alternative method would be wrong. | On an unseen survey bias item, the student should identify a systematic design flaw that sample size cannot repair. The answer must be reached inside the 2.14-minute average without omitting the controlling definition or condition. |
| 9 | experimental design | If this item was slow or wrong, classify the first failed step as condition; do not label the whole result 'careless.' | Repair it by having the student state the exact randomness, independence, count, or distribution check that controls the method. Then solve a contrasting example in which a tempting alternative method would be wrong. | On an unseen experimental design item, the student should separate random assignment for causation from random sampling for generalization. The answer must be reached inside the 2.14-minute average without omitting the controlling definition or condition. |
| 10 | blocking or pairing | If this item was slow or wrong, classify the first failed step as calculator entry; do not label the whole result 'careless.' | Repair it by having the student estimate magnitude first, write the intended function and parameters, then compare the display with the estimate. Then solve a contrasting example in which a tempting alternative method would be wrong. | On an unseen blocking or pairing item, the student should use the matching structure and do not analyze linked observations as independent. The answer must be reached inside the 2.14-minute average without omitting the controlling definition or condition. |
| 11 | simulation | If this item was slow or wrong, classify the first failed step as algebra or arithmetic; do not label the whole result 'careless.' | Repair it by having the student keep extra digits, track parentheses and signs, and substitute the result back into the original expression. Then solve a contrasting example in which a tempting alternative method would be wrong. | On an unseen simulation item, the student should check the label assignment and recorded statistic before counting repetitions. The answer must be reached inside the 2.14-minute average without omitting the controlling definition or condition. |
| 12 | addition rule | If this item was slow or wrong, classify the first failed step as interpretation; do not label the whole result 'careless.' | Repair it by having the student name the population, units, direction, uncertainty, and design limit in a complete contextual sentence. Then solve a contrasting example in which a tempting alternative method would be wrong. | On an unseen addition rule item, the student should draw or define overlapping events so the intersection is not counted twice. The answer must be reached inside the 2.14-minute average without omitting the controlling definition or condition. |
| 13 | conditional probability | If this item was slow or wrong, classify the first failed step as reading; do not label the whole result 'careless.' | Repair it by having the student underline the population, variable, comparison, and requested direction before looking at choices. Then solve a contrasting example in which a tempting alternative method would be wrong. | On an unseen conditional probability item, the student should restrict the denominator to the group after the conditioning bar. The answer must be reached inside the 2.14-minute average without omitting the controlling definition or condition. |
| 14 | independence | If this item was slow or wrong, classify the first failed step as method selection; do not label the whole result 'careless.' | Repair it by having the student write the parameter and data structure, then name why the competing procedure does not apply. Then solve a contrasting example in which a tempting alternative method would be wrong. | On an unseen independence item, the student should verify with a conditional or multiplication rule rather than intuition. The answer must be reached inside the 2.14-minute average without omitting the controlling definition or condition. |
| 15 | expected value | If this item was slow or wrong, classify the first failed step as condition; do not label the whole result 'careless.' | Repair it by having the student state the exact randomness, independence, count, or distribution check that controls the method. Then solve a contrasting example in which a tempting alternative method would be wrong. | On an unseen expected value item, the student should define net outcomes and compute the probability-weighted long-run average. The answer must be reached inside the 2.14-minute average without omitting the controlling definition or condition. |
| 16 | binomial conditions | If this item was slow or wrong, classify the first failed step as calculator entry; do not label the whole result 'careless.' | Repair it by having the student estimate magnitude first, write the intended function and parameters, then compare the display with the estimate. Then solve a contrasting example in which a tempting alternative method would be wrong. | On an unseen binomial conditions item, the student should verify fixed n, two outcomes, independence, and constant p. The answer must be reached inside the 2.14-minute average without omitting the controlling definition or condition. |
| 17 | binomial calculation | If this item was slow or wrong, classify the first failed step as algebra or arithmetic; do not label the whole result 'careless.' | Repair it by having the student keep extra digits, track parentheses and signs, and substitute the result back into the original expression. Then solve a contrasting example in which a tempting alternative method would be wrong. | On an unseen binomial calculation item, the student should look for an efficient complement before entering a long probability sum. The answer must be reached inside the 2.14-minute average without omitting the controlling definition or condition. |
| 18 | normal area | If this item was slow or wrong, classify the first failed step as interpretation; do not label the whole result 'careless.' | Repair it by having the student name the population, units, direction, uncertainty, and design limit in a complete contextual sentence. Then solve a contrasting example in which a tempting alternative method would be wrong. | On an unseen normal area item, the student should sketch the requested region and confirm left, right, or middle probability. The answer must be reached inside the 2.14-minute average without omitting the controlling definition or condition. |
| 19 | normal percentile | If this item was slow or wrong, classify the first failed step as reading; do not label the whole result 'careless.' | Repair it by having the student underline the population, variable, comparison, and requested direction before looking at choices. Then solve a contrasting example in which a tempting alternative method would be wrong. | On an unseen normal percentile item, the student should find z first, then return to the original scale and units. The answer must be reached inside the 2.14-minute average without omitting the controlling definition or condition. |
| 20 | sampling distribution of p-hat | If this item was slow or wrong, classify the first failed step as method selection; do not label the whole result 'careless.' | Repair it by having the student write the parameter and data structure, then name why the competing procedure does not apply. Then solve a contrasting example in which a tempting alternative method would be wrong. | On an unseen sampling distribution of p-hat item, the student should distinguish p, p-hat, and p0 while checking large counts. The answer must be reached inside the 2.14-minute average without omitting the controlling definition or condition. |
| 21 | one-proportion interval | If this item was slow or wrong, classify the first failed step as condition; do not label the whole result 'careless.' | Repair it by having the student state the exact randomness, independence, count, or distribution check that controls the method. Then solve a contrasting example in which a tempting alternative method would be wrong. | On an unseen one-proportion interval item, the student should use observed counts and interpret confidence as long-run method performance. The answer must be reached inside the 2.14-minute average without omitting the controlling definition or condition. |
| 22 | proportion sample size | If this item was slow or wrong, classify the first failed step as calculator entry; do not label the whole result 'careless.' | Repair it by having the student estimate magnitude first, write the intended function and parameters, then compare the display with the estimate. Then solve a contrasting example in which a tempting alternative method would be wrong. | On an unseen proportion sample size item, the student should use a planning value, compute, and round upward. The answer must be reached inside the 2.14-minute average without omitting the controlling definition or condition. |
| 23 | one-proportion test | If this item was slow or wrong, classify the first failed step as algebra or arithmetic; do not label the whole result 'careless.' | Repair it by having the student keep extra digits, track parentheses and signs, and substitute the result back into the original expression. Then solve a contrasting example in which a tempting alternative method would be wrong. | On an unseen one-proportion test item, the student should use p0 in the null standard error and the prespecified tail. The answer must be reached inside the 2.14-minute average without omitting the controlling definition or condition. |
| 24 | two-proportion interval | If this item was slow or wrong, classify the first failed step as interpretation; do not label the whole result 'careless.' | Repair it by having the student name the population, units, direction, uncertainty, and design limit in a complete contextual sentence. Then solve a contrasting example in which a tempting alternative method would be wrong. | On an unseen two-proportion interval item, the student should define p1-p2 and use separate sample estimates in the standard error. The answer must be reached inside the 2.14-minute average without omitting the controlling definition or condition. |
| 25 | two-proportion test | If this item was slow or wrong, classify the first failed step as reading; do not label the whole result 'careless.' | Repair it by having the student underline the population, variable, comparison, and requested direction before looking at choices. Then solve a contrasting example in which a tempting alternative method would be wrong. | On an unseen two-proportion test item, the student should pool successes only because the null sets the population proportions equal. The answer must be reached inside the 2.14-minute average without omitting the controlling definition or condition. |
| 26 | chi-square expected count | If this item was slow or wrong, classify the first failed step as method selection; do not label the whole result 'careless.' | Repair it by having the student write the parameter and data structure, then name why the competing procedure does not apply. Then solve a contrasting example in which a tempting alternative method would be wrong. | On an unseen chi-square expected count item, the student should multiply marginal totals and divide by the grand total. The answer must be reached inside the 2.14-minute average without omitting the controlling definition or condition. |
| 27 | chi-square conclusion | If this item was slow or wrong, classify the first failed step as condition; do not label the whole result 'careless.' | Repair it by having the student state the exact randomness, independence, count, or distribution check that controls the method. Then solve a contrasting example in which a tempting alternative method would be wrong. | On an unseen chi-square conclusion item, the student should say association, not linear correlation or causation. The answer must be reached inside the 2.14-minute average without omitting the controlling definition or condition. |
| 28 | sampling distribution of x-bar | If this item was slow or wrong, classify the first failed step as calculator entry; do not label the whole result 'careless.' | Repair it by having the student estimate magnitude first, write the intended function and parameters, then compare the display with the estimate. Then solve a contrasting example in which a tempting alternative method would be wrong. | On an unseen sampling distribution of x-bar item, the student should use sigma over square root n and justify the normal approximation. The answer must be reached inside the 2.14-minute average without omitting the controlling definition or condition. |
| 29 | t distribution | If this item was slow or wrong, classify the first failed step as algebra or arithmetic; do not label the whole result 'careless.' | Repair it by having the student keep extra digits, track parentheses and signs, and substitute the result back into the original expression. Then solve a contrasting example in which a tempting alternative method would be wrong. | On an unseen t distribution item, the student should recognize unknown sigma, use s, and track degrees of freedom. The answer must be reached inside the 2.14-minute average without omitting the controlling definition or condition. |
| 30 | one-mean interval | If this item was slow or wrong, classify the first failed step as interpretation; do not label the whole result 'careless.' | Repair it by having the student name the population, units, direction, uncertainty, and design limit in a complete contextual sentence. Then solve a contrasting example in which a tempting alternative method would be wrong. | On an unseen one-mean interval item, the student should check randomness, independence, shape, and outliers before interpreting. The answer must be reached inside the 2.14-minute average without omitting the controlling definition or condition. |
| 31 | one-mean test | If this item was slow or wrong, classify the first failed step as reading; do not label the whole result 'careless.' | Repair it by having the student underline the population, variable, comparison, and requested direction before looking at choices. Then solve a contrasting example in which a tempting alternative method would be wrong. | On an unseen one-mean test item, the student should write hypotheses about mu and state evidence at the chosen alpha. The answer must be reached inside the 2.14-minute average without omitting the controlling definition or condition. |
| 32 | paired t procedure | If this item was slow or wrong, classify the first failed step as method selection; do not label the whole result 'careless.' | Repair it by having the student write the parameter and data structure, then name why the competing procedure does not apply. Then solve a contrasting example in which a tempting alternative method would be wrong. | On an unseen paired t procedure item, the student should form one difference per pair and check the difference distribution. The answer must be reached inside the 2.14-minute average without omitting the controlling definition or condition. |
| 33 | two-mean t procedure | If this item was slow or wrong, classify the first failed step as condition; do not label the whole result 'careless.' | Repair it by having the student state the exact randomness, independence, count, or distribution check that controls the method. Then solve a contrasting example in which a tempting alternative method would be wrong. | On an unseen two-mean t procedure item, the student should keep independent groups separate and do not force equal variances. The answer must be reached inside the 2.14-minute average without omitting the controlling definition or condition. |
| 34 | Type I and II errors | If this item was slow or wrong, classify the first failed step as calculator entry; do not label the whole result 'careless.' | Repair it by having the student estimate magnitude first, write the intended function and parameters, then compare the display with the estimate. Then solve a contrasting example in which a tempting alternative method would be wrong. | On an unseen Type I and II errors item, the student should translate each wrong decision into the context before choosing. The answer must be reached inside the 2.14-minute average without omitting the controlling definition or condition. |
| 35 | power | If this item was slow or wrong, classify the first failed step as algebra or arithmetic; do not label the whole result 'careless.' | Repair it by having the student keep extra digits, track parentheses and signs, and substitute the result back into the original expression. Then solve a contrasting example in which a tempting alternative method would be wrong. | On an unseen power item, the student should connect effect size, n, variability, and alpha without changing alpha after data. The answer must be reached inside the 2.14-minute average without omitting the controlling definition or condition. |
| 36 | scatterplot | If this item was slow or wrong, classify the first failed step as interpretation; do not label the whole result 'careless.' | Repair it by having the student name the population, units, direction, uncertainty, and design limit in a complete contextual sentence. Then solve a contrasting example in which a tempting alternative method would be wrong. | On an unseen scatterplot item, the student should describe direction, form, strength, and unusual points before using r. The answer must be reached inside the 2.14-minute average without omitting the controlling definition or condition. |
| 37 | correlation | If this item was slow or wrong, classify the first failed step as reading; do not label the whole result 'careless.' | Repair it by having the student underline the population, variable, comparison, and requested direction before looking at choices. Then solve a contrasting example in which a tempting alternative method would be wrong. | On an unseen correlation item, the student should interpret linear strength without causal language or sensitivity blindness. The answer must be reached inside the 2.14-minute average without omitting the controlling definition or condition. |
| 38 | regression slope | If this item was slow or wrong, classify the first failed step as method selection; do not label the whole result 'careless.' | Repair it by having the student write the parameter and data structure, then name why the competing procedure does not apply. Then solve a contrasting example in which a tempting alternative method would be wrong. | On an unseen regression slope item, the student should state predicted response change per one explanatory unit with units. The answer must be reached inside the 2.14-minute average without omitting the controlling definition or condition. |
| 39 | residual | If this item was slow or wrong, classify the first failed step as condition; do not label the whole result 'careless.' | Repair it by having the student state the exact randomness, independence, count, or distribution check that controls the method. Then solve a contrasting example in which a tempting alternative method would be wrong. | On an unseen residual item, the student should compute observed minus predicted and interpret the sign. The answer must be reached inside the 2.14-minute average without omitting the controlling definition or condition. |
| 40 | r-squared | If this item was slow or wrong, classify the first failed step as calculator entry; do not label the whole result 'careless.' | Repair it by having the student estimate magnitude first, write the intended function and parameters, then compare the display with the estimate. Then solve a contrasting example in which a tempting alternative method would be wrong. | On an unseen r-squared item, the student should name the response variation explained by the fitted linear relationship. The answer must be reached inside the 2.14-minute average without omitting the controlling definition or condition. |
| 41 | influence | If this item was slow or wrong, classify the first failed step as algebra or arithmetic; do not label the whole result 'careless.' | Repair it by having the student keep extra digits, track parentheses and signs, and substitute the result back into the original expression. Then solve a contrasting example in which a tempting alternative method would be wrong. | On an unseen influence item, the student should distinguish high leverage, large residual, and change after refitting. The answer must be reached inside the 2.14-minute average without omitting the controlling definition or condition. |
| 42 | mixed reasoning | If this item was slow or wrong, classify the first failed step as interpretation; do not label the whole result 'careless.' | Repair it by having the student name the population, units, direction, uncertainty, and design limit in a complete contextual sentence. Then solve a contrasting example in which a tempting alternative method would be wrong. | On an unseen mixed reasoning item, the student should identify the parameter, design, method, and allowable conclusion before computing. The answer must be reached inside the 2.14-minute average without omitting the controlling definition or condition. |
A four-question free-response pacing plan
College Board gives 90 minutes for the entire section. It does not require equal time per question. The following 20/20/25/25 split is a practical rehearsal plan based on the revised question roles; it is not an official per-question rule.
| Question | Suggested rehearsal time | Primary role | Time use |
|---|---|---|---|
| 1 | 20 minutes | Formulate questions and collect data | Read all parts; define population/variables; describe the design and its scope; reserve two minutes to check randomization language |
| 2 | 20 minutes | Analyze data and interpret results | Select displays or summaries, calculate what is requested, and connect every value to the context |
| 3 | 25 minutes | Inference | Define the parameter; state method and conditions; compute; write a contextual interval/test conclusion |
| 4 | 25 minutes | Multiple content areas | Scan every part, answer independent entry points, carry results forward, and finish with integrated interpretation |
An alternative is to reserve five minutes at the end by targeting 19, 19, 23, and 24 minutes. The correct plan is the one tested under realistic practice and adjusted from evidence. A student who consistently earns more on inference may allocate differently from one who needs extra reading time on study design.
Within each FRQ, read all parts before calculating. Later parts may be answerable even when an earlier value is missing. Show method, conditions, and setup so the reasoning remains visible. Typed responses should use short complete sentences rather than one dense paragraph that hides the requested conclusions.
Twenty-four FRQ timing components
The suggested ranges fit inside the broader 20/20/25/25 rehearsal plan. They show where time creates points: definitions, design, conditions, calculations, and conclusions.
| Component | Typical rehearsal range | Action | Completion evidence |
|---|---|---|---|
| Q1: formulate the question | 2-3 minutes | Write population, cases, variables, and a comparison or relationship that anticipates variation. | The question determines what data would answer it and avoids undefined words such as better without a measure. |
| Q1: identify the design | 2-3 minutes | Name observational study, sample survey, or experiment and identify selection and assignment mechanisms. | Sampling and assignment are not conflated; the source of data is explicit. |
| Q1: propose random sampling | 3-4 minutes | Define the frame and chance process, including strata or clusters when relevant. | Every selected case has a design-based chance and the represented population is honest. |
| Q1: propose random assignment | 3-4 minutes | State treatments, comparison, random assignment, controls, replication, and any block or pair structure. | The causal contrast is isolated rather than mixed with time, teacher, dose, or participant choice. |
| Q1: evaluate bias or confounding | 3-4 minutes | Name the specific systematic threat and explain the direction or mechanism when possible. | The response does not claim that a larger sample automatically removes bias. |
| Q1: state scope | 2-3 minutes | Connect random sampling to generalization and random assignment to causation. | The conclusion is no broader than the actual frame and implemented treatment. |
| Q2: select a display | 2-3 minutes | Match categorical or quantitative variables to a display and choose a common scale for comparisons. | The display reveals the feature named in the question and does not distort group differences. |
| Q2: calculate summaries | 3-4 minutes | Compute appropriate center, spread, proportion, standardized value, residual, or other requested statistic. | Units and denominators are correct, and calculator output is not pasted without meaning. |
| Q2: describe a distribution | 3 minutes | Address shape, center, spread, unusual features, and context; use comparative language for groups. | The description agrees with the graph and uses resistant summaries when skew/outliers demand them. |
| Q2: compare evidence | 3-4 minutes | Tie numerical and graphical differences to the investigative question. | The response distinguishes observed sample differences from population or causal conclusions. |
| Q2: diagnose a model | 3-4 minutes | Inspect residuals, normal evidence, expected counts, or simulation behavior as the method requires. | A systematic pattern or failed condition is named rather than hidden behind a summary statistic. |
| Q2: interpret | 3 minutes | Translate the result into cases, variables, units, direction, and limitations. | The answer says what the number means and avoids claims the design cannot support. |
| Q3: define parameter/hypotheses | 2-3 minutes | Name p, mu, a difference, or mean paired difference in the population; put equality in H0. | Notation and prose identify the same population quantity and alternative direction. |
| Q3: name procedure | 1 minute | Choose the interval or test matching variable type, groups, and design. | One-proportion, two-proportion, one-mean, paired, two-mean, or chi-square logic is explicit. |
| Q3: check randomness and independence | 2-3 minutes | Use the stated random process and 10 percent condition when sampling without replacement. | The check is linked to the actual population size and sample design rather than assumed. |
| Q3: check counts or shape | 2-3 minutes | Use observed/null large counts for proportions, expected counts for chi-square, or skew/outlier reasoning for t procedures. | The correct method-specific condition is stated with numbers from the prompt. |
| Q3: calculate | 5-7 minutes | Show estimate, standard error, statistic or margin, probability, and interval/test result. | Pooled versus unpooled formulas, tails, degrees of freedom, and rounding are correct. |
| Q3: conclude | 3-4 minutes | Write confidence or evidence language naming population, parameter, direction, and context. | A p-value is not the probability H0 is true and an interval does not contain a percentage of observations. |
| Q4: scan all parts | 2 minutes | Mark independent entry points, supplied values, and parts that can be answered even if an earlier calculation is missing. | The time plan follows available work rather than visual paragraph length. |
| Q4: organize scratch work | 2 minutes | Label parts and keep definitions, values, and formulas adjacent to the corresponding answer. | Typed responses can be transferred without losing a sign, denominator, or symbol definition. |
| Q4: connect design to analysis | 4-5 minutes | Use how data were produced to choose analysis and limit interpretation. | The response does not treat a convenience sample or observational association as randomized evidence. |
| Q4: combine units | 7-9 minutes | Move among distributions, probability, inference, or regression while preserving one contextual thread. | Each method answers its part and intermediate results are carried forward transparently. |
| Q4: evaluate a claim | 3-4 minutes | Compare statistical evidence with the exact wording of the claim and identify practical or design limits. | The conclusion separates significance, effect size, and causation. |
| Q4: final audit | 2-3 minutes | Check every part for a response, units, context, condition, and conclusion. | No accessible point is lost because a later sentence was omitted after correct arithmetic. |
A five-minute timeline for the 180 testing minutes
This is a practice overlay, not an official script. It gives students a way to see drift early and protects an attempt on every question.
| Window | Section | Primary action | Checkpoint |
|---|---|---|---|
| Minute 0-5 | Section I: MCQ | Move from roughly Question 1 through Question 2. Read the statistical task before the story details, calculate only what the item requests, and flag a true blocker. | At the block end, confirm every reached item has a choice and note whether drift came from reading, method selection, or calculator entry. |
| Minute 5-10 | Section I: MCQ | Move from roughly Question 3 through Question 5. Read the statistical task before the story details, calculate only what the item requests, and flag a true blocker. | At the block end, confirm every reached item has a choice and note whether drift came from reading, method selection, or calculator entry. |
| Minute 10-15 | Section I: MCQ | Move from roughly Question 6 through Question 7. Read the statistical task before the story details, calculate only what the item requests, and flag a true blocker. | At the block end, confirm every reached item has a choice and note whether drift came from reading, method selection, or calculator entry. |
| Minute 15-20 | Section I: MCQ | Move from roughly Question 8 through Question 9. Read the statistical task before the story details, calculate only what the item requests, and flag a true blocker. | At the block end, confirm every reached item has a choice and note whether drift came from reading, method selection, or calculator entry. |
| Minute 20-25 | Section I: MCQ | Move from roughly Question 10 through Question 12. Read the statistical task before the story details, calculate only what the item requests, and flag a true blocker. | At the block end, confirm every reached item has a choice and note whether drift came from reading, method selection, or calculator entry. |
| Minute 25-30 | Section I: MCQ | Move from roughly Question 13 through Question 14. Read the statistical task before the story details, calculate only what the item requests, and flag a true blocker. | At the block end, confirm every reached item has a choice and note whether drift came from reading, method selection, or calculator entry. |
| Minute 30-35 | Section I: MCQ | Move from roughly Question 15 through Question 16. Read the statistical task before the story details, calculate only what the item requests, and flag a true blocker. | At the block end, confirm every reached item has a choice and note whether drift came from reading, method selection, or calculator entry. |
| Minute 35-40 | Section I: MCQ | Move from roughly Question 17 through Question 19. Read the statistical task before the story details, calculate only what the item requests, and flag a true blocker. | At the block end, confirm every reached item has a choice and note whether drift came from reading, method selection, or calculator entry. |
| Minute 40-45 | Section I: MCQ | Move from roughly Question 20 through Question 21. Read the statistical task before the story details, calculate only what the item requests, and flag a true blocker. | At the block end, confirm every reached item has a choice and note whether drift came from reading, method selection, or calculator entry. |
| Minute 45-50 | Section I: MCQ | Move from roughly Question 22 through Question 23. Read the statistical task before the story details, calculate only what the item requests, and flag a true blocker. | At the block end, confirm every reached item has a choice and note whether drift came from reading, method selection, or calculator entry. |
| Minute 50-55 | Section I: MCQ | Move from roughly Question 24 through Question 26. Read the statistical task before the story details, calculate only what the item requests, and flag a true blocker. | At the block end, confirm every reached item has a choice and note whether drift came from reading, method selection, or calculator entry. |
| Minute 55-60 | Section I: MCQ | Move from roughly Question 27 through Question 28. Read the statistical task before the story details, calculate only what the item requests, and flag a true blocker. | At the block end, confirm every reached item has a choice and note whether drift came from reading, method selection, or calculator entry. |
| Minute 60-65 | Section I: MCQ | Move from roughly Question 29 through Question 30. Read the statistical task before the story details, calculate only what the item requests, and flag a true blocker. | At the block end, confirm every reached item has a choice and note whether drift came from reading, method selection, or calculator entry. |
| Minute 65-70 | Section I: MCQ | Move from roughly Question 31 through Question 33. Read the statistical task before the story details, calculate only what the item requests, and flag a true blocker. | At the block end, confirm every reached item has a choice and note whether drift came from reading, method selection, or calculator entry. |
| Minute 70-75 | Section I: MCQ | Move from roughly Question 34 through Question 35. Read the statistical task before the story details, calculate only what the item requests, and flag a true blocker. | At the block end, confirm every reached item has a choice and note whether drift came from reading, method selection, or calculator entry. |
| Minute 75-80 | Section I: MCQ | Move from roughly Question 36 through Question 37. Read the statistical task before the story details, calculate only what the item requests, and flag a true blocker. | At the block end, confirm every reached item has a choice and note whether drift came from reading, method selection, or calculator entry. |
| Minute 80-85 | Section I: MCQ | Move from roughly Question 38 through Question 40. Read the statistical task before the story details, calculate only what the item requests, and flag a true blocker. | At the block end, confirm every reached item has a choice and note whether drift came from reading, method selection, or calculator entry. |
| Minute 85-90 | Section I: MCQ | Move from roughly Question 41 through Question 42. Read the statistical task before the story details, calculate only what the item requests, and flag a true blocker. | At the block end, confirm every reached item has a choice and note whether drift came from reading, method selection, or calculator entry. |
| Minute 90-95 | Section II: FRQ | Work on Question 1, whose rehearsal role is to formulate questions and collect data. Read all parts, answer independent entry points, and keep calculation and conclusion visible. | At the block end, compare actual progress with the 20/20/25/25 plan and move rather than polishing one part at the expense of later questions. |
| Minute 95-100 | Section II: FRQ | Work on Question 1, whose rehearsal role is to formulate questions and collect data. Read all parts, answer independent entry points, and keep calculation and conclusion visible. | At the block end, compare actual progress with the 20/20/25/25 plan and move rather than polishing one part at the expense of later questions. |
| Minute 100-105 | Section II: FRQ | Work on Question 1, whose rehearsal role is to formulate questions and collect data. Read all parts, answer independent entry points, and keep calculation and conclusion visible. | At the block end, compare actual progress with the 20/20/25/25 plan and move rather than polishing one part at the expense of later questions. |
| Minute 105-110 | Section II: FRQ | Work on Question 1, whose rehearsal role is to formulate questions and collect data. Read all parts, answer independent entry points, and keep calculation and conclusion visible. | At the block end, compare actual progress with the 20/20/25/25 plan and move rather than polishing one part at the expense of later questions. |
| Minute 110-115 | Section II: FRQ | Work on Question 2, whose rehearsal role is to analyze data and interpret results. Read all parts, answer independent entry points, and keep calculation and conclusion visible. | At the block end, compare actual progress with the 20/20/25/25 plan and move rather than polishing one part at the expense of later questions. |
| Minute 115-120 | Section II: FRQ | Work on Question 2, whose rehearsal role is to analyze data and interpret results. Read all parts, answer independent entry points, and keep calculation and conclusion visible. | At the block end, compare actual progress with the 20/20/25/25 plan and move rather than polishing one part at the expense of later questions. |
| Minute 120-125 | Section II: FRQ | Work on Question 2, whose rehearsal role is to analyze data and interpret results. Read all parts, answer independent entry points, and keep calculation and conclusion visible. | At the block end, compare actual progress with the 20/20/25/25 plan and move rather than polishing one part at the expense of later questions. |
| Minute 125-130 | Section II: FRQ | Work on Question 2, whose rehearsal role is to analyze data and interpret results. Read all parts, answer independent entry points, and keep calculation and conclusion visible. | At the block end, compare actual progress with the 20/20/25/25 plan and move rather than polishing one part at the expense of later questions. |
| Minute 130-135 | Section II: FRQ | Work on Question 3, whose rehearsal role is to complete inference. Read all parts, answer independent entry points, and keep calculation and conclusion visible. | At the block end, compare actual progress with the 20/20/25/25 plan and move rather than polishing one part at the expense of later questions. |
| Minute 135-140 | Section II: FRQ | Work on Question 3, whose rehearsal role is to complete inference. Read all parts, answer independent entry points, and keep calculation and conclusion visible. | At the block end, compare actual progress with the 20/20/25/25 plan and move rather than polishing one part at the expense of later questions. |
| Minute 140-145 | Section II: FRQ | Work on Question 3, whose rehearsal role is to complete inference. Read all parts, answer independent entry points, and keep calculation and conclusion visible. | At the block end, compare actual progress with the 20/20/25/25 plan and move rather than polishing one part at the expense of later questions. |
| Minute 145-150 | Section II: FRQ | Work on Question 3, whose rehearsal role is to complete inference. Read all parts, answer independent entry points, and keep calculation and conclusion visible. | At the block end, compare actual progress with the 20/20/25/25 plan and move rather than polishing one part at the expense of later questions. |
| Minute 150-155 | Section II: FRQ | Work on Question 3, whose rehearsal role is to complete inference. Read all parts, answer independent entry points, and keep calculation and conclusion visible. | At the block end, compare actual progress with the 20/20/25/25 plan and move rather than polishing one part at the expense of later questions. |
| Minute 155-160 | Section II: FRQ | Work on Question 4, whose rehearsal role is to connect multiple content areas. Read all parts, answer independent entry points, and keep calculation and conclusion visible. | At the block end, compare actual progress with the 20/20/25/25 plan and move rather than polishing one part at the expense of later questions. |
| Minute 160-165 | Section II: FRQ | Work on Question 4, whose rehearsal role is to connect multiple content areas. Read all parts, answer independent entry points, and keep calculation and conclusion visible. | At the block end, compare actual progress with the 20/20/25/25 plan and move rather than polishing one part at the expense of later questions. |
| Minute 165-170 | Section II: FRQ | Work on Question 4, whose rehearsal role is to connect multiple content areas. Read all parts, answer independent entry points, and keep calculation and conclusion visible. | At the block end, compare actual progress with the 20/20/25/25 plan and move rather than polishing one part at the expense of later questions. |
| Minute 170-175 | Section II: FRQ | Work on Question 4, whose rehearsal role is to connect multiple content areas. Read all parts, answer independent entry points, and keep calculation and conclusion visible. | At the block end, compare actual progress with the 20/20/25/25 plan and move rather than polishing one part at the expense of later questions. |
| Minute 175-180 | Section II: FRQ | Work on Question 4, whose rehearsal role is to connect multiple content areas. Read all parts, answer independent entry points, and keep calculation and conclusion visible. | At the block end, compare actual progress with the 20/20/25/25 plan and move rather than polishing one part at the expense of later questions. |
Twenty timing decisions and recoveries
| Situation | Best decision | Recovery principle |
|---|---|---|
| After 15 minutes, only 5 MCQs are complete. | The target is about 7 questions. Shorten rereading, use a skip-return decision, and avoid spending several minutes proving one option when another item may be faster. | Recover gradually over the next two seven-question blocks; do not rush every remaining question indiscriminately. |
| After 30 minutes, 16 MCQs are complete. | The target is about 14, so the student has a small cushion. Maintain the same reasoning quality rather than deliberately slowing down. | Use the cushion for a flagged shared-prompt set or a careful final pass. |
| A probability set shares one long prompt across three items. | Read the setup once, define events and assumptions on scratch paper, and reuse that representation. Do not restart the context for every item. | If one part blocks progress, the other two may still be answerable from the common information. |
| A calculator probability does not match any option. | Estimate the expected magnitude, check tail direction, parameters, and complement use. One diagnostic pass is worth more than repeated identical entry. | If the mismatch remains, eliminate impossible options, mark the item, and return after completing the section. |
| An unfamiliar word appears in an otherwise standard question. | Use the definitions and relationships shown in the prompt. AP questions often provide enough structure to reason without prior knowledge of the story context. | Translate the story into variable type, parameter, and requested statistic before deciding it is unfamiliar. |
| Two choices differ only in denominator. | Stop and name the reference group. Conditional probability, survey proportions, and rates often test the denominator more than arithmetic. | Write the numerator and denominator in words before dividing. |
| The student reaches Question 28 at 68 minutes. | The target for Question 28 is 60 minutes, so 14 questions remain with 22 minutes. Use a disciplined first pass on every remaining item. | Protect time to touch all questions because unanswered items and wrong answers both earn no credit. |
| A regression question invites a long calculation. | Check whether the question asks for interpretation of slope, residual, or r-squared that can be read directly from output. | Use the meaning of the requested quantity before reconstructing the entire regression. |
| A question has already consumed four minutes. | Unless completion is immediate, record useful work, select the best supported option, flag it, and move. Four minutes is almost twice the average budget. | Return only after the remainder of the section has been seen. |
| Ten minutes remain with six unanswered MCQs. | Use about one minute per item for an informed first pass, then spend the remaining minutes on the strongest flagged opportunities. | Do not devote all ten minutes to the first difficult item and leave five unseen. |
| The student finishes MCQs with eight minutes left. | Review marked items, calculator entries, signs, tails, and denominators. Change an answer only for a specific statistical reason. | Confirm every question has a recorded choice before the section ends. |
| FRQ 1 consumes 30 minutes. | A personal 20-minute target has been exceeded by ten minutes. Stop polishing and move so later questions can earn points. | Recover by preserving an attempt on every requested part, not by deleting all checking time. |
| An FRQ part cannot be solved numerically. | State the method, define the parameter, check conditions, show the setup, and interpret what can be determined. Later parts may be independent. | Partial statistical reasoning can earn credit; a blank response cannot. |
| A condition fails during inference. | Do not force the standard procedure. State the failed condition and explain how it limits the requested calculation or conclusion. | Answer later conceptual parts with the failure carried forward honestly. |
| An FRQ asks for a justification, not a calculation. | Write the design, condition, or statistical principle that proves the claim. Extra calculator output does not replace the requested reason. | Use complete contextual sentences and stop once the justification is sufficient. |
| Question 4 has many subparts and 24 minutes remain. | Scan the complete question, identify independent entry points, and allocate time by available points or tasks rather than by visual length. | Answer accessible parts first, then return to linked calculations. |
| A typed formula is hard to format. | Use clear conventional keyboard notation, define symbols, and place the calculation on a readable line. Mathematical meaning matters more than decorative typesetting. | Use the Bluebook symbols menu when helpful and keep prose conclusions separate from formulas. |
| The calculator and hand calculation disagree slightly. | Check whether one value was rounded too early. Keep extra digits through the standard error, test statistic, or probability, then round the final answer sensibly. | A small rounding difference is different from a wrong tail, wrong model, or wrong denominator. |
| Five minutes remain in FRQ with one conclusion unwritten. | Write the conclusion first: population, parameter or difference, direction, strength of evidence or interval, and design limitation if requested. | A final contextual sentence can secure interpretation credit even when earlier arithmetic is imperfect. |
| A student finishes far early on every practice set but accuracy is low. | The problem is not lack of speed. Add a deliberate method-selection and condition check before committing to an answer. | Use saved time to verify definitions, signs, tails, denominators, and contextual wording. |
Sixteen timed drills
Use the drills in order. Short work isolates one pacing skill; full sections then test whether the skill survives mixed content and digital response entry.
| Drill | Time | Purpose | Completion standard |
|---|---|---|---|
| Drill 1: 14 MCQs | 30 minutes | One-third section at exact average pace | Record time at Q7 and Q14; classify every delay by reading, reasoning, calculator, or indecision. |
| Drill 2: 21 MCQs | 45 minutes | Half section with mixed units | Use one skip-return pass and ensure all 21 receive an answer before review. |
| Drill 3: probability set | 12 minutes | One shared prompt plus three items and two independent probability questions | Build one event model and reuse it without mixing conditional denominators. |
| Drill 4: regression set | 12 minutes | One shared regression prompt plus interpretation and residual items | Read output once; identify what can be answered directly before calculating. |
| Drill 5: 42 MCQs | 90 minutes | First full revised section | Use checkpoints at Q14=30 minutes and Q28=60 minutes; preserve an answer on every item. |
| Drill 6: FRQ 1 | 20 minutes | Formulate a question and evaluate data collection | Define population and variables, describe random mechanisms, and state allowable scope. |
| Drill 7: FRQ 2 | 20 minutes | Analyze data and interpret results | Connect displays and summaries to a contextual conclusion rather than listing output. |
| Drill 8: FRQ 3 | 25 minutes | Inference | Write parameter, method, conditions, calculation, and conclusion in a fixed sequence. |
| Drill 9: FRQ 4 | 25 minutes | Multi-focus investigation | Scan all parts, answer independent entry points, and carry earlier results forward clearly. |
| Drill 10: four FRQs | 90 minutes | Full revised free-response section | Use a 20/20/25/25 rehearsal plan, then compare actual time and points by question. |
| Drill 11: digital notation | 35 minutes | Retype three solved responses from scratch without copying | Use unambiguous p-hat, x-bar, H0/Ha, inequality, interval, and conclusion forms. |
| Drill 12: calculator recovery | 30 minutes | Ten deliberately varied calculator tasks | For every mismatch, estimate magnitude and diagnose list, tail, parameter, or mode error. |
| Drill 13: condition sprint | 20 minutes | Name conditions for 15 methods or designs | Conditions must be method-specific; 'n is large' alone is not accepted. |
| Drill 14: conclusion sprint | 20 minutes | Rewrite 12 incomplete conclusions | Each conclusion names population, parameter, context, direction, and uncertainty/evidence. |
| Drill 15: full exam | 180 minutes testing plus the same planned break/transition routine | 42 MCQs followed by four FRQs | Measure section performance separately and audit whether fatigue changes method selection or writing. |
| Drill 16: error repair | 60 minutes | Redo every miss from Drill 15 without notes | Explain why the original response failed and what cue should trigger the correct method next time. |
Timing a fully digital AP Statistics workflow
The revised exam places both multiple-choice and free-response work in Bluebook. Students still use scratch paper for planning and calculations and may use an approved handheld calculator in addition to built-in Desmos. Practice should therefore measure the full loop: read on screen, sketch or calculate, consult reference information when needed, and enter an unambiguous answer.
Do not make the first full timed set the first Bluebook-style rehearsal. Practice account access, navigation, keyboard notation, symbols, calculator placement, and scratch-paper organization separately. Once those mechanics are stable, timed scores better represent statistical reasoning rather than interface unfamiliarity.
A strong scratch page labels question numbers and keeps essential values near the corresponding work. For FRQs, write the parameter, conditions, and core calculation before typing. This reduces the risk of losing a conclusion because the calculation is scattered or the symbol definitions are forgotten.
Use the digital exam guide for device setup and the format guide for question structure. Keeping detailed setup outside this timing page lets pacing remain the main purpose.
Timing with approved accommodations
Approved accommodations can alter timing, breaks, setting, or response method. The student’s schedule comes from the approved plan and school instructions. A generic percentage extension cannot safely predict reporting, transition, break, or dismissal details.
Practice should match the authorized arrangement as closely as the school recommends. If extended time is approved, use it strategically rather than treating every question as permission for unlimited rereading. Cumulative checkpoints can be scaled to the authorized section time, while the same skip-return and complete-every-part principles remain useful.
Questions about eligibility, approval status, or implementation belong with the school’s SSD coordinator or AP coordinator and College Board’s official accommodations process. This page describes the standard 90/90 structure, not an individual’s approved schedule.
Frequently asked questions about AP Statistics exam length
How long is the AP Statistics Exam?
The revised exam provides three hours of testing: 90 minutes for multiple choice and 90 minutes for free response.
Does the three hours include the break?
The three hours are section testing time. Administrative procedures and any scheduled break make the total room commitment longer.
How many MCQs are completed in 90 minutes?
The revised section contains 42 questions. The mathematical average is about 2.14 minutes per question, but actual item times should vary.
How long is each FRQ?
College Board sets a 90-minute section total for four questions, not a universal minute limit for each question. A 20/20/25/25 split is one rehearsal plan, not an official rule.
Are both sections worth the same amount?
Yes. Multiple choice contributes 50% and free response contributes 50% of the exam score.
Does a shared prompt add extra time?
No separate time is added. Two MCQ sets contain three questions sharing a prompt, so efficient reuse of the setup matters.
Should every MCQ receive exactly 2.14 minutes?
No. Easy interpretation items may take under a minute, while a multistep probability question may take longer. Use cumulative checkpoints.
Is guessing penalized?
College Board’s general timing page states that the MCQ score is based on correct answers and that wrong or unanswered answers do not lose extra points. Use an informed answer rather than leaving an item blank.
Can a calculator save time?
Yes when the method and entry are already practiced. Searching menus or repeating an unexplained mismatch can consume more time than the calculation saves.
What if a student has approved extended time?
Use the individualized schedule supplied by the school. Do not apply a generic multiplier to predict arrival, break, or dismissal details.
How long should a full practice exam take?
Use 180 minutes of section testing and reproduce a realistic transition and break routine. Review time happens after the simulation, not inside it.
What is the best MCQ checkpoint?
Question 14 at about 30 minutes and Question 28 at about 60 minutes divide the section into thirds. More frequent seven-question checks can diagnose drift.
What if one FRQ is unfinished?
Move to accessible parts of later questions before the section ends. Method, conditions, setup, and interpretation can earn credit even when a numerical step is incomplete.
Does fully digital delivery shorten the exam?
No. The revised assessment still lists 90 minutes per section. Digital delivery changes response mechanics, not the stated testing total.
When should timing practice begin?
Begin short checkpoints after methods are understood, then add half sections, full sections, and finally a full exam. Speed built on incorrect methods is not useful.
Official sources
| Official source | What it verifies |
|---|---|
| AP Statistics assessment | Three-hour duration, 42 MCQs in 90 minutes, four FRQs in 90 minutes, and 50/50 weighting |
| AP Statistics revisions | Revised question counts and fully digital May 2027 delivery |
| AP Exam timing and structure | General MCQ scoring and timing guidance |
| Bluebook | Official digital testing application and practice access |
| AP calculator policy | Calculator rules for exam preparation and testing |
