Yes. Beginning with the May 2027 administration, AP Statistics is fully digital. Students complete both 42 multiple-choice questions and all four free-response questions in Bluebook. They still receive scratch paper and printed reference information, can access the reference digitally, and may use an approved handheld calculator in addition to built-in Desmos.
Is the AP Statistics Exam digital?
For May 2027, the entire standard AP Statistics Exam is completed in Bluebook. There is no paper free-response booklet. Fully digital describes the response and submission method; it does not mean an unproctored at-home test.
The exact 2027 subject date was still not displayed on the checked public calendar as of July 18, 2026, but the digital transition itself is explicitly confirmed for May 2027. Date and delivery method are separate facts.
What changes from the 2026 hybrid exam
| Feature | 2026 hybrid administration | Beginning May 2027 |
|---|---|---|
| Multiple choice | Completed in Bluebook | Completed in Bluebook |
| Free response | Viewed digitally but answered in paper booklets | Viewed and answered in Bluebook |
| FRQ booklet | Paper response booklet | No paper FRQ booklet for the fully digital administration |
| Scratch work | Paper available under exam procedures | Scratch paper continues for planning and calculations |
| Reference information | Subject reference materials | Printed reference booklet continues and reference is also digital |
| Calculator | Statistical calculator expected | Built-in Desmos plus approved handheld option |
| Notation | Handwritten FRQ notation | Typed notation with symbols menu and digital-response guidance |
Bluebook setup that should be completed early
Use the official Bluebook site or the school’s managed-device process. Current system requirements, installation steps, and practice access belong to College Board and the school, not a frozen device list on a blog. The exact device intended for testing should complete the readiness and practice process.
Students should know their College Board account credentials and practice entering them. A saved browser or password-manager entry may not be available in the testing environment. Duplicate accounts should be resolved with the school and official support rather than multiplied with another new account.
Official practice or preview should be used to learn navigation, answer review, response fields, symbols, calculator access, and the submission flow. Interface familiarity is not content mastery, so practice must also include aligned AP Statistics reasoning.
Device readiness checklist
| Area | Check | Evidence |
|---|---|---|
| Ownership/management | Confirm whether the school supplies, manages, or approves the intended device | Coordinator or technical-support confirmation |
| Operating system | Compare the live Bluebook requirements with the exact installed version | Current readiness check passes |
| Application | Install/update through the official or managed process | Bluebook opens and official practice loads |
| Account | Enter credentials manually and resolve duplicate-account issues | Correct account sign-in succeeds |
| Keyboard | Test letters, numbers, punctuation, inequalities, editing, and navigation | A long response types without missing/repeated characters |
| Pointing device | Test trackpad or approved mouse under the expected desk setup | Selections and cursor placement are reliable |
| Display | Use supported scale/accessibility settings and confirm all controls are visible | Prompt, response, and tools are readable without overlap |
| Battery | Run a long practice and inspect drain | Safe capacity or authorized power plan |
| Charger | Verify condition, connector, and school rules | Device charges normally |
| Network | Test in the school/managed environment | Practice works under relevant filtering and authentication |
| Calculator | Verify policy, batteries, and primary workflow | Correct output on mixed statistical tasks |
| Support | Record school technical and coordinator contacts | A known path exists for pre-exam issues |
The digital statistics workflow: screen, scratch, calculator, response
Read on screen. Identify the command verb, population, variables, design, and requested conclusion before moving to arithmetic. In a shared-prompt MCQ set, define the common context once and preserve it across all three items.
Plan on scratch paper. Label the question and part, define symbols, sketch graphs or tails, write conditions, and estimate expected magnitude. Scratch work reduces transfer errors only when it is organized.
Calculate with a practiced tool. Use built-in Desmos or an approved handheld. Write the intended probability, standard error, interval, test, or regression quantity before entry; then compare output with the estimate.
Type scoring evidence. Enter enough method, conditions, setup, and calculation to make reasoning visible. Follow with a complete contextual conclusion. Do not type every abandoned scratch step or omit interpretation because the calculator produced a value.
Review. Check signs, tails, denominators, pooled versus unpooled formulas, degrees of freedom, units, notation definitions, and whether every subpart has a response.
Fifty clear typed-notation examples
The symbols menu can help, but conventional keyboard notation is effective when it is defined and grouped unambiguously. The examples below show meaning rather than one mandatory style.
| Purpose | Readable entry | Use | Avoid |
|---|---|---|---|
| Population proportion | Define p in words before using it in hypotheses or a conclusion. | Do not use p for an observed sample proportion. | |
| Sample proportion | Use consistent keyboard notation and state x/n when helpful. | Do not drop the hat and silently turn a statistic into a parameter. | |
| Population mean | Define the population and quantitative variable with units. | Do not label the sample mean mu. | |
| Sample mean | Type x-bar consistently and attach the original units in interpretation. | Do not use x for the mean unless it is explicitly defined. | |
| Population SD | Use for known population spread or model statements. | Do not replace an unknown sigma with a guessed value. | |
| Sample SD | Use s when the spread is calculated from sample data. | Do not confuse s with standard error. | |
| Null hypothesis | Equality belongs in H0 and symbols must match the defined parameter. | Do not write H0 about p-hat. | |
| Alternative hypothesis | Choose >, <, or != from the research question before seeing results. | Do not change direction after observing the sample. | |
| Complement | Define event A and use a superscript or clear text such as not A. | Do not subtract an unrelated event probability. | |
| Union | Use or for the union and account for overlap with the general addition rule. | Do not assume events are disjoint without evidence. | |
| Intersection | Use and for simultaneous occurrence and multiply conditionally when needed. | Do not multiply marginal probabilities unless independence applies. | |
| Conditional probability | State in words that the denominator is restricted to B. | Do not reverse P(A|B) and P(B|A). | |
| z-score | Use parentheses so the entire numerator is divided by sigma. | Do not type x – mu/sigma without grouping. | |
| Normal probability | Use an inequality and confirm the tail with a quick sketch on scratch paper. | Do not report a left-tail calculator value for a right-tail question. | |
| Binomial model | Define X as the number of successes and verify all model conditions. | Do not use the model for a changing p or dependent trials without justification. | |
| Binomial probability | Parenthesize 1-p and make exponent scope clear. | Do not omit the combination count. | |
| Expected value | Define outcomes as net values when a game has a cost. | Do not describe E(X) as a guaranteed single result. | |
| Sample-mean SE | Separate the spread of sample means from the spread of individual observations. | Do not divide sigma by n. | |
| Sample-proportion SE | Use the appropriate p source for probability, interval, or null test. | Do not interchange p, p-hat, and p0. | |
| Margin of error | Show the critical value and standard error before the product. | Do not call the full interval width the margin of error. | |
| Confidence interval | Type the final ordered endpoints and interpret the population parameter. | Do not say the interval contains a percentage of raw observations. | |
| One-proportion test | Use p0 throughout the null standard error and group the denominator. | Do not use p-hat in the null SE. | |
| One-mean t | Show the nested denominator clearly and include df=n-1. | Do not omit parentheses around s/sqrt(n). | |
| Paired difference | Define the subtraction order and use d-bar, s_d, and n pairs. | Do not treat the two measurement columns as independent groups. | |
| Two proportions | Name group 1 and group 2 before interpreting signs. | Do not reverse order in the conclusion. | |
| Pooled proportion | Use only under the equality null in a two-proportion z test. | Do not pool an ordinary confidence interval. | |
| Expected count | Label the cell and preserve marginal totals. | Do not use the observed cell count in place of E. | |
| Chi-square statistic | Use the symbols menu or readable text and state degrees of freedom. | Do not infer direction from squared contributions alone. | |
| Degrees of freedom | Use for a chi-square table with r rows and c columns. | Do not use r*c or the grand total. | |
| p-value | Write the conditioning phrase in words when symbolic entry is awkward. | Do not type P(H0 is true). | |
| Reject decision | Follow with evidence for Ha in context. | Do not claim the result proves Ha with certainty. | |
| Fail-to-reject decision | State insufficient evidence for Ha in context. | Do not accept H0 as exactly true. | |
| Regression equation | Define x and y, include units in slope interpretation, and keep the hat on predicted y. | Do not call y-hat an observed response. | |
| Residual | Positive means observed is above predicted. | Do not reverse the subtraction. | |
| Correlation | Interpret direction and strength of linear association. | Do not use r as a slope or causal effect. | |
| Coefficient of determination | Write that 64% of sample response variation is explained by the fitted linear model. | Do not say 64% of points lie on the line. | |
| Slope interpretation | Name both variables and their units and stay within the observed range. | Do not claim causation without random assignment. | |
| Confidence interpretation | Name the population parameter and interval endpoints with units. | Do not assign 95% probability after the interval is computed. | |
| Simulation estimate | Define a success and the simulated trial. | Do not claim repetitions repair an invalid random assignment. | |
| At least | Include k and use a complement when efficient. | Do not enter P(X > k). | |
| At most | Include k and check calculator bounds. | Do not use a strict inequality. | |
| Not equal | Use the symbols menu or type != consistently. | Do not write a one-sided conclusion for a two-sided test. | |
| Square root | Use brackets or parentheses to show the full radicand. | Do not take the square root of only p. | |
| Exponent | Parenthesize the base and exponent when plain text could be ambiguous. | Do not type 1-p^n-k. | |
| Interval endpoints | Use ordered endpoints and explain whether zero or a null value lies inside. | Do not confuse parentheses with the parameter's probability. | |
| Approximation | Use approximately when a model or rounded computation is used. | Do not display more digits than the data and method support. | |
| Units | Attach units to means, SDs, slopes, residuals, and interval endpoints. | Do not label a proportion difference as percent when percentage points are meant. | |
| Parameter definition | Write this before hypotheses or inference conclusions. | Do not leave population and trait implicit. | |
| Condition statement | Use actual values and the correct interval/test source of p. | Do not type only large counts are met. | |
| Scope statement | Name which random mechanisms are present in the study. | Do not infer both from the word random alone. |
Fifty original typed-notation drills
Use these in Bluebook practice or a plain response field. Each drill requires mathematical entry plus the statistical sentence that gives it meaning.
| Skill | Drill | Target entry | Completion standard |
|---|---|---|---|
| Population proportion | Create an original one-line response that requires p. Define the symbols first, place parentheses around every grouped numerator or denominator, and follow the line with one contextual sentence. | The response must satisfy this meaning: Define p in words before using it in hypotheses or a conclusion. It fails if the student does the following: Do not use p for an observed sample proportion. | |
| Sample proportion | On blank scratch paper, write what sample proportion means; then enter p-hat in a response field without copying. Read the entry aloud to test whether its scope and operators are unambiguous. | The response must satisfy this meaning: Use consistent keyboard notation and state x/n when helpful. It fails if the student does the following: Do not drop the hat and silently turn a statistic into a parameter. | |
| Population mean | Use a hypothetical sample or model to produce mu. Keep enough digits for the computation, round the final value sensibly, and state the population or random variable the line concerns. | The response must satisfy this meaning: Define the population and quantitative variable with units. It fails if the student does the following: Do not label the sample mean mu. | |
| Sample mean | Diagnose a deliberately ambiguous version of x-bar, repair the grouping or symbol definition, and type the corrected version. Explain why the repair changes meaning rather than appearance alone. | The response must satisfy this meaning: Type x-bar consistently and attach the original units in interpretation. It fails if the student does the following: Do not use x for the mean unless it is explicitly defined. | |
| Population SD | Combine sigma with a complete AP-style explanation: identify the method or principle, enter the mathematical line, and finish with the result's meaning and design limitation. | The response must satisfy this meaning: Use for known population spread or model statements. It fails if the student does the following: Do not replace an unknown sigma with a guessed value. | |
| Sample SD | Create an original one-line response that requires s. Define the symbols first, place parentheses around every grouped numerator or denominator, and follow the line with one contextual sentence. | The response must satisfy this meaning: Use s when the spread is calculated from sample data. It fails if the student does the following: Do not confuse s with standard error. | |
| Null hypothesis | On blank scratch paper, write what null hypothesis means; then enter H0: p = 0.60 in a response field without copying. Read the entry aloud to test whether its scope and operators are unambiguous. | The response must satisfy this meaning: Equality belongs in H0 and symbols must match the defined parameter. It fails if the student does the following: Do not write H0 about p-hat. | |
| Alternative hypothesis | Use a hypothetical sample or model to produce Ha: p > 0.60. Keep enough digits for the computation, round the final value sensibly, and state the population or random variable the line concerns. | The response must satisfy this meaning: Choose >, <, or != from the research question before seeing results. It fails if the student does the following: Do not change direction after observing the sample. | |
| Complement | Diagnose a deliberately ambiguous version of P(A^c) = 1 – P(A), repair the grouping or symbol definition, and type the corrected version. Explain why the repair changes meaning rather than appearance alone. | The response must satisfy this meaning: Define event A and use a superscript or clear text such as not A. It fails if the student does the following: Do not subtract an unrelated event probability. | |
| Union | Combine P(A or B) with a complete AP-style explanation: identify the method or principle, enter the mathematical line, and finish with the result's meaning and design limitation. | The response must satisfy this meaning: Use or for the union and account for overlap with the general addition rule. It fails if the student does the following: Do not assume events are disjoint without evidence. | |
| Intersection | Create an original one-line response that requires P(A and B). Define the symbols first, place parentheses around every grouped numerator or denominator, and follow the line with one contextual sentence. | The response must satisfy this meaning: Use and for simultaneous occurrence and multiply conditionally when needed. It fails if the student does the following: Do not multiply marginal probabilities unless independence applies. | |
| Conditional probability | On blank scratch paper, write what conditional probability means; then enter P(A | B) in a response field without copying. Read the entry aloud to test whether its scope and operators are unambiguous. | The response must satisfy this meaning: State in words that the denominator is restricted to B. It fails if the student does the following: Do not reverse P(A|B) and P(B|A). | |
| z-score | Use a hypothetical sample or model to produce z = (x – mu) / sigma. Keep enough digits for the computation, round the final value sensibly, and state the population or random variable the line concerns. | The response must satisfy this meaning: Use parentheses so the entire numerator is divided by sigma. It fails if the student does the following: Do not type x – mu/sigma without grouping. | |
| Normal probability | Diagnose a deliberately ambiguous version of P(Z > 1.50), repair the grouping or symbol definition, and type the corrected version. Explain why the repair changes meaning rather than appearance alone. | The response must satisfy this meaning: Use an inequality and confirm the tail with a quick sketch on scratch paper. It fails if the student does the following: Do not report a left-tail calculator value for a right-tail question. | |
| Binomial model | Combine X ~ Binomial(n, p) with a complete AP-style explanation: identify the method or principle, enter the mathematical line, and finish with the result's meaning and design limitation. | The response must satisfy this meaning: Define X as the number of successes and verify all model conditions. It fails if the student does the following: Do not use the model for a changing p or dependent trials without justification. | |
| Binomial probability | Create an original one-line response that requires P(X = k) = C(n,k)p^k(1-p)^(n-k). Define the symbols first, place parentheses around every grouped numerator or denominator, and follow the line with one contextual sentence. | The response must satisfy this meaning: Parenthesize 1-p and make exponent scope clear. It fails if the student does the following: Do not omit the combination count. | |
| Expected value | On blank scratch paper, write what expected value means; then enter E(X) = sum[x_i p_i] in a response field without copying. Read the entry aloud to test whether its scope and operators are unambiguous. | The response must satisfy this meaning: Define outcomes as net values when a game has a cost. It fails if the student does the following: Do not describe E(X) as a guaranteed single result. | |
| Sample-mean SE | Use a hypothetical sample or model to produce SE(x-bar) = sigma / sqrt(n). Keep enough digits for the computation, round the final value sensibly, and state the population or random variable the line concerns. | The response must satisfy this meaning: Separate the spread of sample means from the spread of individual observations. It fails if the student does the following: Do not divide sigma by n. | |
| Sample-proportion SE | Diagnose a deliberately ambiguous version of SE(p-hat) = sqrt[p(1-p)/n], repair the grouping or symbol definition, and type the corrected version. Explain why the repair changes meaning rather than appearance alone. | The response must satisfy this meaning: Use the appropriate p source for probability, interval, or null test. It fails if the student does the following: Do not interchange p, p-hat, and p0. | |
| Margin of error | Combine ME = critical value * SE with a complete AP-style explanation: identify the method or principle, enter the mathematical line, and finish with the result's meaning and design limitation. | The response must satisfy this meaning: Show the critical value and standard error before the product. It fails if the student does the following: Do not call the full interval width the margin of error. | |
| Confidence interval | Create an original one-line response that requires estimate +/- critical value * SE. Define the symbols first, place parentheses around every grouped numerator or denominator, and follow the line with one contextual sentence. | The response must satisfy this meaning: Type the final ordered endpoints and interpret the population parameter. It fails if the student does the following: Do not say the interval contains a percentage of raw observations. | |
| One-proportion test | On blank scratch paper, write what one-proportion test means; then enter z = (p-hat – p0) / sqrt[p0(1-p0)/n] in a response field without copying. Read the entry aloud to test whether its scope and operators are unambiguous. | The response must satisfy this meaning: Use p0 throughout the null standard error and group the denominator. It fails if the student does the following: Do not use p-hat in the null SE. | |
| One-mean t | Use a hypothetical sample or model to produce t = (x-bar – mu0) / (s/sqrt(n)). Keep enough digits for the computation, round the final value sensibly, and state the population or random variable the line concerns. | The response must satisfy this meaning: Show the nested denominator clearly and include df=n-1. It fails if the student does the following: Do not omit parentheses around s/sqrt(n). | |
| Paired difference | Diagnose a deliberately ambiguous version of d = after – before, repair the grouping or symbol definition, and type the corrected version. Explain why the repair changes meaning rather than appearance alone. | The response must satisfy this meaning: Define the subtraction order and use d-bar, s_d, and n pairs. It fails if the student does the following: Do not treat the two measurement columns as independent groups. | |
| Two proportions | Combine parameter: p1 – p2 with a complete AP-style explanation: identify the method or principle, enter the mathematical line, and finish with the result's meaning and design limitation. | The response must satisfy this meaning: Name group 1 and group 2 before interpreting signs. It fails if the student does the following: Do not reverse order in the conclusion. | |
| Pooled proportion | Create an original one-line response that requires p-pool = (x1+x2)/(n1+n2). Define the symbols first, place parentheses around every grouped numerator or denominator, and follow the line with one contextual sentence. | The response must satisfy this meaning: Use only under the equality null in a two-proportion z test. It fails if the student does the following: Do not pool an ordinary confidence interval. | |
| Expected count | On blank scratch paper, write what expected count means; then enter E = (row total)(column total)/(grand total) in a response field without copying. Read the entry aloud to test whether its scope and operators are unambiguous. | The response must satisfy this meaning: Label the cell and preserve marginal totals. It fails if the student does the following: Do not use the observed cell count in place of E. | |
| Chi-square statistic | Use a hypothetical sample or model to produce chi-square = sum[(O-E)^2/E]. Keep enough digits for the computation, round the final value sensibly, and state the population or random variable the line concerns. | The response must satisfy this meaning: Use the symbols menu or readable text and state degrees of freedom. It fails if the student does the following: Do not infer direction from squared contributions alone. | |
| Degrees of freedom | Diagnose a deliberately ambiguous version of df = (r-1)(c-1), repair the grouping or symbol definition, and type the corrected version. Explain why the repair changes meaning rather than appearance alone. | The response must satisfy this meaning: Use for a chi-square table with r rows and c columns. It fails if the student does the following: Do not use r*c or the grand total. | |
| p-value | Combine p-value = P(statistic at least as extreme | H0) with a complete AP-style explanation: identify the method or principle, enter the mathematical line, and finish with the result's meaning and design limitation. | The response must satisfy this meaning: Write the conditioning phrase in words when symbolic entry is awkward. It fails if the student does the following: Do not type P(H0 is true). | |
| Reject decision | Create an original one-line response that requires p-value < alpha, so reject H0. Define the symbols first, place parentheses around every grouped numerator or denominator, and follow the line with one contextual sentence. | The response must satisfy this meaning: Follow with evidence for Ha in context. It fails if the student does the following: Do not claim the result proves Ha with certainty. | |
| Fail-to-reject decision | On blank scratch paper, write what fail-to-reject decision means; then enter p-value > alpha, so fail to reject H0 in a response field without copying. Read the entry aloud to test whether its scope and operators are unambiguous. | The response must satisfy this meaning: State insufficient evidence for Ha in context. It fails if the student does the following: Do not accept H0 as exactly true. | |
| Regression equation | Use a hypothetical sample or model to produce y-hat = b0 + b1x. Keep enough digits for the computation, round the final value sensibly, and state the population or random variable the line concerns. | The response must satisfy this meaning: Define x and y, include units in slope interpretation, and keep the hat on predicted y. It fails if the student does the following: Do not call y-hat an observed response. | |
| Residual | Diagnose a deliberately ambiguous version of residual = y – y-hat, repair the grouping or symbol definition, and type the corrected version. Explain why the repair changes meaning rather than appearance alone. | The response must satisfy this meaning: Positive means observed is above predicted. It fails if the student does the following: Do not reverse the subtraction. | |
| Correlation | Combine -1 <= r <= 1 with a complete AP-style explanation: identify the method or principle, enter the mathematical line, and finish with the result's meaning and design limitation. | The response must satisfy this meaning: Interpret direction and strength of linear association. It fails if the student does the following: Do not use r as a slope or causal effect. | |
| Coefficient of determination | Create an original one-line response that requires r^2 = 0.64. Define the symbols first, place parentheses around every grouped numerator or denominator, and follow the line with one contextual sentence. | The response must satisfy this meaning: Write that 64% of sample response variation is explained by the fitted linear model. It fails if the student does the following: Do not say 64% of points lie on the line. | |
| Slope interpretation | On blank scratch paper, write what slope interpretation means; then enter For each +1 unit in x, predicted y changes by b1 units in a response field without copying. Read the entry aloud to test whether its scope and operators are unambiguous. | The response must satisfy this meaning: Name both variables and their units and stay within the observed range. It fails if the student does the following: Do not claim causation without random assignment. | |
| Confidence interpretation | Use a hypothetical sample or model to produce We are 95% confident that …. Keep enough digits for the computation, round the final value sensibly, and state the population or random variable the line concerns. | The response must satisfy this meaning: Name the population parameter and interval endpoints with units. It fails if the student does the following: Do not assign 95% probability after the interval is computed. | |
| Simulation estimate | Diagnose a deliberately ambiguous version of estimated P(event) = successes / repetitions, repair the grouping or symbol definition, and type the corrected version. Explain why the repair changes meaning rather than appearance alone. | The response must satisfy this meaning: Define a success and the simulated trial. It fails if the student does the following: Do not claim repetitions repair an invalid random assignment. | |
| At least | Combine P(X >= k) with a complete AP-style explanation: identify the method or principle, enter the mathematical line, and finish with the result's meaning and design limitation. | The response must satisfy this meaning: Include k and use a complement when efficient. It fails if the student does the following: Do not enter P(X > k). | |
| At most | Create an original one-line response that requires P(X <= k). Define the symbols first, place parentheses around every grouped numerator or denominator, and follow the line with one contextual sentence. | The response must satisfy this meaning: Include k and check calculator bounds. It fails if the student does the following: Do not use a strict inequality. | |
| Not equal | On blank scratch paper, write what not equal means; then enter Ha: parameter != null value in a response field without copying. Read the entry aloud to test whether its scope and operators are unambiguous. | The response must satisfy this meaning: Use the symbols menu or type != consistently. It fails if the student does the following: Do not write a one-sided conclusion for a two-sided test. | |
| Square root | Use a hypothetical sample or model to produce sqrt[p(1-p)/n]. Keep enough digits for the computation, round the final value sensibly, and state the population or random variable the line concerns. | The response must satisfy this meaning: Use brackets or parentheses to show the full radicand. It fails if the student does the following: Do not take the square root of only p. | |
| Exponent | Diagnose a deliberately ambiguous version of (1-p)^(n-k), repair the grouping or symbol definition, and type the corrected version. Explain why the repair changes meaning rather than appearance alone. | The response must satisfy this meaning: Parenthesize the base and exponent when plain text could be ambiguous. It fails if the student does the following: Do not type 1-p^n-k. | |
| Interval endpoints | Combine (lower, upper) with a complete AP-style explanation: identify the method or principle, enter the mathematical line, and finish with the result's meaning and design limitation. | The response must satisfy this meaning: Use ordered endpoints and explain whether zero or a null value lies inside. It fails if the student does the following: Do not confuse parentheses with the parameter's probability. | |
| Approximation | Create an original one-line response that requires approximately 0.084. Define the symbols first, place parentheses around every grouped numerator or denominator, and follow the line with one contextual sentence. | The response must satisfy this meaning: Use approximately when a model or rounded computation is used. It fails if the student does the following: Do not display more digits than the data and method support. | |
| Units | On blank scratch paper, write what units means; then enter minutes, grams, points, or percentage points in a response field without copying. Read the entry aloud to test whether its scope and operators are unambiguous. | The response must satisfy this meaning: Attach units to means, SDs, slopes, residuals, and interval endpoints. It fails if the student does the following: Do not label a proportion difference as percent when percentage points are meant. | |
| Parameter definition | Use a hypothetical sample or model to produce Let p be the true proportion of … in …. Keep enough digits for the computation, round the final value sensibly, and state the population or random variable the line concerns. | The response must satisfy this meaning: Write this before hypotheses or inference conclusions. It fails if the student does the following: Do not leave population and trait implicit. | |
| Condition statement | Diagnose a deliberately ambiguous version of n*p0 = … and n*(1-p0) = …, both >= 10, repair the grouping or symbol definition, and type the corrected version. Explain why the repair changes meaning rather than appearance alone. | The response must satisfy this meaning: Use actual values and the correct interval/test source of p. It fails if the student does the following: Do not type only large counts are met. | |
| Scope statement | Combine Random assignment supports causation; random sampling supports generalization with a complete AP-style explanation: identify the method or principle, enter the mathematical line, and finish with the result's meaning and design limitation. | The response must satisfy this meaning: Name which random mechanisms are present in the study. It fails if the student does the following: Do not infer both from the word random alone. |
Built-in Desmos and approved handheld calculators
College Board says students may use an approved handheld graphing calculator in addition to the built-in Desmos graphing calculator in Bluebook. The calculator policy expects statistical capabilities and permits up to two approved handhelds under the current rules. Verify the live policy rather than relying on a static list.
Choose a primary workflow for lists and summaries, normal and binomial probabilities, confidence intervals and tests, chi-square tables, and regression. A second calculator is a backup, not a second unpracticed menu system. If Desmos is the primary tool, practice the current AP Statistics functions in official preview conditions.
Technology does not choose a parameter, justify independence, decide pooled versus unpooled standard error, or interpret a p-value. Write the statistical setup first and estimate the answer’s direction and scale. That habit detects many entry errors before time is lost.
Scratch paper and reference information on a digital exam
Fully digital does not mean paperless. College Board’s revision FAQ says students continue to receive scratch paper and a printed reference information booklet, and the same reference information is available through Bluebook. Follow the proctor’s handling directions for both.
Use scratch paper for diagrams, event definitions, distribution sketches, condition checks, formulas, calculator inputs, and intermediate values. Number every block. At the end of a part, mark the final value and the contextual conclusion that must appear on screen.
Use the reference booklet to verify exact formulas and tables, but know the method map. A formula sheet cannot decide whether data are paired, whether the null SE is pooled, whether a chi-square expected-count condition holds, or whether random assignment permits causation.
Thirty digital troubleshooting cases
| Problem | Risk | Response | Resolved when | Prevention |
|---|---|---|---|---|
| Bluebook is not installed | No verified testing application is available on the intended device. | Use the official Bluebook download/setup path or the school's managed-device process well before exam day. | Successful sign-in and launch of official practice/preview on the intended hardware. | Prevent recurrence by scheduling the same check on the exact intended device after every major update and again during the final week. |
| Unsupported operating system | The device may not meet current Bluebook requirements. | Check the live official device page and coordinate an approved school device; do not rely on an old requirements screenshot. | The exact device passes the current readiness check. | Document the verified setup and owner of the next action so the problem is not rediscovered on exam morning. |
| School-managed installation is blocked | The student may lack permission to install or update applications. | Contact school technical support and the AP coordinator instead of changing management settings independently. | Bluebook appears in the managed account and opens under school policy. | Repeat the repair under school conditions, because a home-only success does not establish readiness in the managed environment. |
| Password works only through saved autofill | The testing environment may not provide the saved credential. | Practice entering the College Board account information and complete official recovery before exam week. | Manual sign-in succeeds without exposing the password in notes or messages. | Add this item to the preflight checklist and require observable evidence rather than a verbal assumption that it should work. |
| Duplicate College Board accounts | Course enrollment and testing identity may be split. | Work through the school and official account support; do not create another account as a shortcut. | The coordinator confirms the correct account and AP registration association. | Rehearse the statistical task immediately after the technical repair so interface readiness and content readiness are both confirmed. |
| Keyboard key fails | Typed FRQ notation or prose could become unreliable. | Repair or replace the device through the authorized plan; test letters, numbers, punctuation, and navigation keys. | A full practice response can be typed without missed or repeated characters. | Prevent recurrence by scheduling the same check on the exact intended device after every major update and again during the final week. |
| Trackpad or mouse is unreliable | Navigation and selection may consume time or create wrong clicks. | Use an approved functioning pointing method and test it under the expected desk setup. | Menus, choices, scrolling, and text placement work consistently. | Document the verified setup and owner of the next action so the problem is not rediscovered on exam morning. |
| Battery drains quickly | The device may not last through administration and both sections. | Follow the school charging plan, verify charger condition, and report battery health concerns early. | A full-length rehearsal completes with a safe battery margin or authorized power access. | Repeat the repair under school conditions, because a home-only success does not establish readiness in the managed environment. |
| Charger is missing or damaged | The backup power plan is incomplete. | Obtain the correct charger under school policy; do not assume a nearby student's charger is compatible or available. | The device charges normally and cable/adapter are approved for the room. | Add this item to the preflight checklist and require observable evidence rather than a verbal assumption that it should work. |
| Home Wi-Fi works but school network does not | The relevant testing environment remains unverified. | Run school-based readiness/practice and involve school technical support. | Bluebook connects and practice functions on the managed network. | Rehearse the statistical task immediately after the technical repair so interface readiness and content readiness are both confirmed. |
| Application update appears near exam day | An untested update can change readiness or behavior. | Follow official and school update instructions, then rerun the practice check. | Sign-in, navigation, calculator, symbols, and submission practice still work. | Prevent recurrence by scheduling the same check on the exact intended device after every major update and again during the final week. |
| Screen scaling hides controls | Display settings may reduce usable prompt or response area. | Use supported accessibility/display settings and rehearse the exact layout; ask the school before making managed changes. | All controls and text remain readable without overlapping. | Document the verified setup and owner of the next action so the problem is not rediscovered on exam morning. |
| Text cursor jumps | Answers may be inserted in the wrong place. | Test keyboard/trackpad behavior, avoid resting contact that moves the pointer, and obtain device support if persistent. | A multi-paragraph response can be edited accurately. | Repeat the repair under school conditions, because a home-only success does not establish readiness in the managed environment. |
| Symbols menu is unfamiliar | Notation entry can interrupt statistical reasoning. | Open official practice and locate common symbols before timed work; keep readable keyboard alternatives. | p-hat, x-bar, mu, inequalities, and subscripts can be entered or defined efficiently. | Add this item to the preflight checklist and require observable evidence rather than a verbal assumption that it should work. |
| Desmos workflow is unfamiliar | Built-in availability will not save time without practiced commands and interpretation. | Rehearse the statistical calculations allowed by the current tool using official practice contexts. | Results are estimated for reasonableness and translated into statistical conclusions. | Rehearse the statistical task immediately after the technical repair so interface readiness and content readiness are both confirmed. |
| Handheld calculator is unverified | The model or capability may not comply with policy. | Check the current College Board calculator policy and school instructions. | The coordinator/student has verified the model and required statistical functions. | Prevent recurrence by scheduling the same check on the exact intended device after every major update and again during the final week. |
| Calculator batteries are weak | A permitted tool may fail midsection. | Replace/charge batteries and follow the policy for an approved backup calculator. | A full timed practice completes without power warnings. | Document the verified setup and owner of the next action so the problem is not rediscovered on exam morning. |
| Calculator output differs from expectation | The error may be tail, list, bounds, parameter, or mode rather than the device. | Estimate first, inspect inputs once, and compare with a manual setup instead of repeating identical entry. | The corrected output matches direction, scale, and the written formula. | Repeat the repair under school conditions, because a home-only success does not establish readiness in the managed environment. |
| Scratch work has no question labels | Transfer errors become likely when values and conclusions are mixed. | Number every scratch block and reserve a line for the final result to type. | Each on-screen response can be traced to one organized scratch section. | Add this item to the preflight checklist and require observable evidence rather than a verbal assumption that it should work. |
| Reference booklet is ignored | Time is lost reconstructing supplied formulas or tables. | Practice locating the relevant page while still learning what each formula means. | The reference is used selectively without replacing method selection. | Rehearse the statistical task immediately after the technical repair so interface readiness and content readiness are both confirmed. |
| Reference booklet is overused | Constant searching interrupts flow and may hide conceptual gaps. | Memorize meanings and decision cues; consult only for exact forms or tables. | The student can name method and conditions before opening the reference. | Prevent recurrence by scheduling the same check on the exact intended device after every major update and again during the final week. |
| Typed answer contains one dense paragraph | Scoring evidence for conditions, calculation, and conclusion is hard to find. | Use short paragraphs or labeled parts, with equations on separate lines. | A reader can locate parameter, method, conditions, calculation, and conclusion quickly. | Document the verified setup and owner of the next action so the problem is not rediscovered on exam morning. |
| Notation is ambiguous | A scorer cannot tell which quantity or denominator the student means. | Define symbols in words and add parentheses/brackets for plain-text formulas. | Another reader can reproduce the calculation from the typed line. | Repeat the repair under school conditions, because a home-only success does not establish readiness in the managed environment. |
| Student types every scratch calculation | Time is spent transferring work that does not clarify the response. | Type enough setup and reasoning to make the solution visible, then emphasize the requested interpretation. | The response is complete but not a transcript of abandoned arithmetic. | Add this item to the preflight checklist and require observable evidence rather than a verbal assumption that it should work. |
| A response is calculated but not interpreted | The statistical task remains incomplete. | Add population, parameter, direction, uncertainty/evidence, units, and design scope as requested. | The final sentence answers the question in context. | Rehearse the statistical task immediately after the technical repair so interface readiness and content readiness are both confirmed. |
| A condition fails | Forcing a standard method can create invalid inference. | State the failed condition and explain the consequence; answer independent later parts honestly. | The response does not hide the failure behind calculator output. | Prevent recurrence by scheduling the same check on the exact intended device after every major update and again during the final week. |
| Bluebook freezes during practice | The device or application path needs support before exam day. | Record what happened, use official troubleshooting or school support, and repeat the readiness test. | A later full practice completes normally on the authorized device. | Document the verified setup and owner of the next action so the problem is not rediscovered on exam morning. |
| Unexpected issue occurs during the exam | Independent troubleshooting or external communication can violate procedure. | Notify the proctor immediately and follow the authorized recovery instructions. | The student remains in the official testing process and documents nothing externally. | Repeat the repair under school conditions, because a home-only success does not establish readiness in the managed environment. |
| Student wants to photograph a prompt | AP content security and testing rules prohibit unauthorized capture or sharing. | Do not photograph, copy, transmit, or recreate secure exam content. | The device and scratch materials are handled exactly as the proctor directs. | Add this item to the preflight checklist and require observable evidence rather than a verbal assumption that it should work. |
| Student opens another website | Leaving the authorized workflow can violate testing rules and compromise security. | Use only permitted Bluebook tools, supplied references, and approved calculator resources. | All work remains inside the authorized environment. | Rehearse the statistical task immediately after the technical repair so interface readiness and content readiness are both confirmed. |
Twenty-five digital response quality checks
| Evidence | How to type it | Strong result | Failure |
|---|---|---|---|
| Part label | Start with (a), (b), or the prompt's label when a question has several tasks. | A scorer can match every response to the requested part. | One continuous paragraph with no visible part boundaries. |
| Parameter | Define p, mu, a difference, or mean paired difference in the population before inference. | Hypotheses, calculations, and conclusions use the same target. | Symbols appear without population or variable definitions. |
| Variable roles | Name explanatory and response variables for regression or experiments. | Slope, design, and causal language refer to the correct roles. | x and y switch meaning midway through a response. |
| Subtraction order | Write group 1 minus group 2 or after minus before before calculating. | Signs in estimates, intervals, and conclusions have one consistent meaning. | The final conclusion reverses the calculation order. |
| Method name | Name the interval, test, simulation, design, or model when requested or helpful. | The response can be evaluated against the correct conditions and formula. | A calculator result appears without identifying what produced it. |
| Randomness | Cite the actual sampling or assignment mechanism from the prompt. | Scope of inference is tied to how data were produced. | The word random is asserted with no mechanism. |
| Independence | Use study design and the 10 percent condition when sampling without replacement. | The standard error's independence basis is visible. | A repeated-measures design is treated as independent. |
| Large counts | Type the relevant observed or null expected successes and failures. | The reader can verify the proportion approximation. | The response says n is large without the required counts. |
| Shape/outliers | Describe population/sample shape or cite sample size and inspect unusual values for t methods. | Robustness of the t reference is justified. | A small skewed sample with an outlier is accepted automatically. |
| Expected counts | State the expected-count evidence for chi-square work. | The reference distribution is supported by model-based cell counts. | Observed counts are used for the condition. |
| Formula grouping | Use parentheses or brackets around complete numerators, denominators, and square-root contents. | Another reader can reproduce the intended arithmetic. | Plain text changes operator scope. |
| Calculator value | Identify what the displayed value represents: estimate, SE, statistic, df, probability, or endpoint. | Technology output becomes statistical evidence. | An unlabeled line of decimal numbers. |
| Rounding | Carry extra digits through intermediate calculations and round the final answer to supported precision. | Tail probabilities and interval endpoints remain internally consistent. | Early rounding materially changes the decision. |
| Units | Attach the original measurement units to means, SDs, slopes, residuals, and intervals. | The interpretation remains contextual rather than abstract. | A quantitative answer has no measurement scale. |
| Proportion scale | Distinguish decimal proportions, percentages, and percentage-point differences. | A difference such as 0.08 is described accurately as eight percentage points. | A proportion difference is called eight percent without context. |
| p-value meaning | State the tail probability under H0 using the observed statistic or one more extreme. | The evidence measure is interpreted correctly. | The p-value is called the probability H0 is true. |
| Decision | Compare p-value with prespecified alpha and write reject or fail to reject H0. | The formal decision and contextual sentence agree. | Alpha is changed after seeing the data. |
| Confidence | Name the population parameter and long-run confidence interpretation. | The interval endpoints answer the estimation question. | The response says 95 percent of observations are in the interval. |
| Effect magnitude | Report the estimate and uncertainty, then compare with a practical benchmark when asked. | Statistical and practical importance are separate. | A small p-value is treated as a large effect. |
| Causation | Reserve causal language for a valid randomized treatment comparison. | The conclusion matches the experiment's design. | An observational association is called an effect. |
| Generalization | Name the population represented by the random sample and sampling frame. | The conclusion does not exceed the selected population. | Volunteers are treated as representative solely because n is large. |
| Residual | Type residual=observed-predicted and interpret sign. | The model error has the correct direction and response units. | The subtraction is reversed. |
| r-squared | Name the response variation explained by the fitted linear relationship. | The coefficient of determination receives its correct contextual meaning. | It is called the percentage of points on the line. |
| Limitation | State a design, condition, measurement, model, or practical limit relevant to the conclusion. | The strength of the claim remains proportional to evidence. | A generic more research is needed sentence with no identified issue. |
| Final answer | End each requested part with the direct statistical answer in context. | The scorer does not have to infer the conclusion from scratch arithmetic. | Correct calculations stop without answering the question. |
A thirty-day digital rehearsal plan
| Day | Focus | Action | Evidence |
|---|---|---|---|
| Day 30 | Account | Sign in manually and verify the correct College Board account without creating a new duplicate. | Manual login succeeds and enrollment/testing identity is confirmed. |
| Day 29 | Device | Check current supported-device requirements against the exact intended hardware and operating system. | A dated readiness record names the tested device. |
| Day 28 | Installation | Install or open Bluebook through the official or school-managed process. | The application launches under the account that will be used. |
| Day 27 | Keyboard | Type a 300-word response using letters, numbers, parentheses, inequalities, and editing keys. | No missed, repeated, or misplaced characters occur. |
| Day 26 | Navigation | Use official practice to move among prompts, choices, response fields, review states, and available tools. | The student can navigate without losing answer location. |
| Day 25 | Symbols | Enter or clearly define p-hat, x-bar, mu, sigma, inequalities, subscripts, and chi-square. | Twenty notation examples are unambiguous to another reader. |
| Day 24 | Scratch paper | Solve five mixed questions with one labeled scratch block per question, then type only essential work. | Every typed value traces to an organized calculation. |
| Day 23 | Reference | Locate formulas/tables for descriptive statistics, distributions, intervals, tests, and chi-square. | Each target is found quickly and its meaning is explained. |
| Day 22 | Desmos | Practice supported graphing and statistical calculations with reasonableness estimates. | Output matches a written model and contextual interpretation. |
| Day 21 | Handheld | Verify policy and complete the same calculations on the intended approved calculator. | The student has one dependable primary workflow and a lawful backup plan. |
| Day 20 | Power | Run a 90-minute practice from the expected starting charge and inspect battery use. | The device retains a safe margin and the charger plan is confirmed. |
| Day 19 | Network | Use the school environment or official readiness process rather than relying on home connectivity. | Bluebook practice works under managed network conditions. |
| Day 18 | MCQ set | Complete 14 four-choice questions in 30 minutes with one shared-prompt mini-set. | All choices are recorded and navigation does not create errors. |
| Day 17 | MCQ review | Review flags, calculator entries, tails, denominators, and unanswered-item behavior. | Every change has a specific statistical reason. |
| Day 16 | FRQ design | Type one 20-minute study-design response with short labeled paragraphs. | Population, variables, random mechanism, bias, and scope are visible. |
| Day 15 | FRQ analysis | Type one 20-minute data-analysis response using a graph or output. | Calculation and interpretation are separated and connected. |
| Day 14 | FRQ inference | Type a 25-minute inference response from parameter through conclusion. | Method, conditions, calculation, and contextual conclusion are easy to locate. |
| Day 13 | FRQ mixed | Type a 25-minute multi-focus response and answer independent later parts after one deliberate blocked step. | No accessible part is left blank because an earlier calculation failed. |
| Day 12 | Full MCQ | Complete 42 questions in 90 minutes using the intended screen, scratch, and calculator workflow. | Checkpoints at 14 and 28 are within five minutes of target. |
| Day 11 | MCQ repair | Redo every miss without notes and diagnose content, method, condition, entry, or interpretation. | The corrected reason is written, not merely the answer letter. |
| Day 10 | Full FRQ | Complete four typed questions in 90 minutes with a 20/20/25/25 rehearsal split. | Every part contains a substantive attempt and final conclusions are present. |
| Day 9 | FRQ repair | Rewrite only missing parameter, condition, setup, or conclusion components, then recombine one full response. | The revised response is concise and complete. |
| Day 8 | Update check | After any operating-system or Bluebook update, rerun account, navigation, symbols, and calculator checks. | No regression in the intended setup is found. |
| Day 7 | Full simulation | Complete both 90-minute sections with a realistic transition and break routine. | Scores and fatigue errors are recorded by section. |
| Day 6 | Technical repair | Resolve any battery, keyboard, network, login, display, or managed-device issue through official support. | No known unresolved technical problem remains. |
| Day 5 | Notation speed | Retype ten solved statistical arguments from blank response fields without copying. | Conventional notation is fast, consistent, and defined. |
| Day 4 | Calculator/reference | Complete ten mixed calculations while deciding when reference or technology is actually useful. | Method is named before tool use and output magnitude is checked. |
| Day 3 | Short mixed set | Complete seven MCQs and one FRQ conclusion with full quality but low fatigue. | Performance is stable and no new workflow is introduced. |
| Day 2 | Logistics | Verify device, charger, calculator, account, school instructions, and permitted materials. | A single checklist agrees with the coordinator's final notice. |
| Day 1 | Light rehearsal | Review notation, one calculator workflow, one condition set, and one complete conclusion; protect sleep. | The student stops studying with the setup ready and the account known. |
Exam security and personal privacy
Do not photograph, record, copy, transmit, reconstruct, or discuss secure exam content outside authorized rules. Do not open external websites, messaging, notes, or unapproved applications during testing. Follow the proctor and Bluebook workflow for every interruption.
Protect account credentials. Practice manual sign-in without placing a password in shared documents, screenshots, public posts, or calculator notes. Resolve account problems through official recovery and the school rather than asking another person to log in.
After the exam, handle scratch paper, reference materials, devices, and calculator storage exactly as directed. A public preparation page should teach format and statistics without exposing secure operational content.
Digital testing with approved accommodations
Approved accommodations may affect timing, breaks, setting, assistive technology, display, input, or other conditions. The individualized plan and school instructions control the setup. This page cannot calculate or promise a personal schedule.
Practice the approved configuration early enough for technical staff to resolve compatibility issues. Use only settings and tools authorized for the student. If assistive technology changes how notation is entered or read, rehearse complete statistical arguments, not only isolated symbols.
Questions about approval or implementation should go to the school’s SSD/AP coordinator and College Board’s official accommodations process. Device convenience alone does not create an accommodation.
Exam-day digital workflow
- Follow the school’s reporting time, room, device, charger, calculator, and permitted-material instructions.
- Enter the correct College Board account credentials and follow the proctor’s setup directions.
- Check the keyboard, pointing device, display, and calculator only as the proctor permits.
- During MCQ, reuse shared prompts carefully and record a choice for every item.
- During FRQ, label scratch work, type method and conditions, show the calculation, and conclude in context.
- If anything fails, notify the proctor immediately; do not leave Bluebook or communicate externally.
- Complete submission and material handling exactly as directed.
The statistical standard does not change on screen: define the target, use the design, choose the method, check conditions, calculate accurately, and state what the evidence supports.
Screen fatigue should be addressed during full rehearsals, not discovered during the operational exam. Use the approved display setup, practice deliberate reading and brief visual resets only when permitted, and monitor whether late-section errors come from content, navigation, or attention. The remedy must fit the authorized testing environment.
Frequently asked questions about the digital AP Statistics Exam
Is the AP Statistics Exam digital in 2027?
Yes. Beginning in May 2027, both MCQ and FRQ sections are completed and submitted through Bluebook.
Was the 2026 exam fully digital?
The 2026 design was hybrid: MCQs were digital and FRQs used paper booklets. That should not be presented as the 2027 format.
Are paper FRQ booklets used in 2027?
College Board says the exam will no longer include paper free-response booklets for the fully digital administration.
Is scratch paper still provided?
Yes. College Board’s revision FAQ says students continue to receive scratch paper for planning and calculations.
Is a printed reference booklet provided?
Yes, and reference information is also available in Bluebook.
Is Desmos available?
Yes. Built-in Desmos is available in Bluebook for AP Statistics.
Can a student also use a handheld calculator?
Yes, an approved handheld graphing calculator remains allowed in addition to built-in Desmos.
Does a student need to type every calculation?
Type enough method, setup, and reasoning to make the response clear. Scratch paper can hold planning and arithmetic, but the final submitted answer must contain the evidence needed for scoring.
How should x-bar or p-hat be typed?
Use the symbols menu when helpful or conventional keyboard notation such as x-bar and p-hat, then define it in words and use it consistently.
Can a saved password be used?
Students should know and practice entering their College Board credentials; official AP guidance warns that saved passwords may not work on exam day.
What device should a student buy?
Do not purchase based solely on a blog. Follow current Bluebook requirements and the school’s device plan; many schools provide or manage devices.
What if Bluebook does not work?
Resolve problems through official troubleshooting and school support before exam day. During the exam, notify the proctor immediately.
Can a student open external calculator websites?
Use only the authorized Bluebook tools, supplied materials, and approved handheld calculator. Do not leave the testing workflow.
Does fully digital mean the exam is online at home?
No. AP Exams are administered by authorized schools or test centers under proctoring and security procedures.
Does the digital format change the three-hour length?
No. The revised assessment still lists two 90-minute sections. See the timing guide for pacing.
Official sources
| Official source | What it verifies |
|---|---|
| AP Statistics revisions | Fully digital May 2027 delivery, no paper FRQ booklets, Desmos/handheld use, scratch paper, reference booklet, and symbols guidance |
| AP Statistics assessment | Bluebook delivery, 42 MCQs, four FRQs, calculator and reference links |
| Bluebook | Official application, setup information, and practice access |
| AP calculator policy | Current calculator capabilities, built-in Desmos, and approved handheld rules |
| Reference information for AP exams | AP Statistics reference materials |
| What to bring on AP Exam day | Current permitted and required materials |
| AP Statistics Digital Response Guidelines | Official keyboard notation and digital response guidance |
