A chi-square calculation is only one link in a statistical argument. The learner must identify whether the design compares distributions across populations or treatments, or asks about association between categorical variables in one population; state hypotheses in context; compute expected counts under the null model; verify conditions; calculate the statistic and p-value; and make a conclusion whose scope matches the design.
The direct answer is to build the reasoning from the table outward. Expected counts describe what the null model predicts for each cell. The chi-square statistic accumulates squared discrepancies scaled by those expectations. A small p-value is evidence against the null model, but it does not by itself prove causation or explain which mechanism produced the pattern. The original examples align with the AP Statistics Course and Exam Description effective fall 2026 for the May 2027 fully digital exam.
Find your next learning step
A practical AP Statistics 2027 guide to chi-square tests for homogeneity and independence: identify the design, calculate expected counts, verify conditions, read cell contributions and keep conclusions within the evidence.
ROUTE 1 · CHAPTERS 1–3
Understand the task
Identify the task boundary and the first decisions.
ROUTE 2 · CHAPTERS 4–6
Build the mechanism
Work through the process from first principles.
ROUTE 3 · CHAPTERS 7–9
Use evidence precisely
Connect detail to a defensible result.
ROUTE 4 · CHAPTERS 10–12
Diagnose and transfer
Repair causes and test changed contexts.
ROUTE 5 · CHAPTERS 13–15
Practise and verify
Build a loop and keep source boundaries visible.
Full chapter index · Independent learning check · Education parent guide
Full chapter index
Understand the task · Chapters 1–3
Build the mechanism · Chapters 4–6
Use evidence precisely · Chapters 7–9
Diagnose and transfer · Chapters 10–12
Practise and verify · Chapters 13–15
This chapter focuses on start with the investigative question. Name the population, categorical variables and comparison before choosing a procedure.
Mechanism. Name the population, categorical variables and comparison before choosing a procedure. The learner should make the state of the problem visible before using a technique. This turns a remembered rule into a chain of decisions that another reader can inspect. The explanation should identify what information enters the step, what transformation occurs and what result becomes available for the next step.
Worked case. A random sample of students reports travel mode and year group: two categorical variables in one population. After reaching the result, explain why the step was justified and what evidence would have changed it. That second explanation prevents a correct-looking answer from becoming a lucky routine. It also gives a teacher a precise place to test understanding.
Changed context. Replace one important feature of the case while keeping the underlying decision. Ask what remains stable and what must change. Transfer is demonstrated when the learner preserves the reasoning mechanism without copying surface language, numbers or labels from the first example. A useful comparison names both the invariant decision and the changed evidence.
Failure and repair. A two-way table triggers automatic calculation. Match the design in words first. The repair should be used immediately on a near case and then on a farther case. The near case confirms the correction; the farther case checks whether the learner understands the cause rather than the appearance of the original mistake.
Action. Classify scenarios without numbers and state what the null model would mean. A useful feedback note names the decision, cites the evidence and gives the next small task. It does not merely report that the answer was wrong. Over time, this creates an error history that distinguishes a one-off slip from a recurring misconception.
Student, parent and teacher roles. The student owns the explanation and records the error cause. The teacher chooses the smallest task that can expose the cause and checks a changed case. The parent can protect regular practice, ask the learner to explain the decision aloud and avoid supplying a polished answer that hides the learner's current model.
Evidence boundary. This case is illustrative and original. It demonstrates a reasoning process, not a guaranteed examination response or score. The learner should follow current official instructions and keep each claim within what the prompt, text, design or algorithm can support.
CHAPTER 2 OF 15 · Understand the task
Distinguish homogeneity and independence
This chapter focuses on distinguish homogeneity and independence. Homogeneity compares one categorical distribution across populations or treatments; independence studies two variables in one population.
Mechanism. Homogeneity compares one categorical distribution across populations or treatments; independence studies two variables in one population. The learner should make the state of the problem visible before using a technique. This turns a remembered rule into a chain of decisions that another reader can inspect. The explanation should identify what information enters the step, what transformation occurs and what result becomes available for the next step.
Worked case. Separate town samples reporting ticket type imply homogeneity; one commuter sample reporting town and type implies independence. After reaching the result, explain why the step was justified and what evidence would have changed it. That second explanation prevents a correct-looking answer from becoming a lucky routine. It also gives a teacher a precise place to test understanding.
Changed context. Replace one important feature of the case while keeping the underlying decision. Ask what remains stable and what must change. Transfer is demonstrated when the learner preserves the reasoning mechanism without copying surface language, numbers or labels from the first example. A useful comparison names both the invariant decision and the changed evidence.
Failure and repair. The table alone is used to decide. Read how observations entered it. The repair should be used immediately on a near case and then on a farther case. The near case confirms the correction; the farther case checks whether the learner understands the cause rather than the appearance of the original mistake.
Action. Rewrite one context into the other while holding counts fixed. A useful feedback note names the decision, cites the evidence and gives the next small task. It does not merely report that the answer was wrong. Over time, this creates an error history that distinguishes a one-off slip from a recurring misconception.
Student, parent and teacher roles. The student owns the explanation and records the error cause. The teacher chooses the smallest task that can expose the cause and checks a changed case. The parent can protect regular practice, ask the learner to explain the decision aloud and avoid supplying a polished answer that hides the learner's current model.
Evidence boundary. This case is illustrative and original. It demonstrates a reasoning process, not a guaranteed examination response or score. The learner should follow current official instructions and keep each claim within what the prompt, text, design or algorithm can support.
This chapter focuses on state contextual hypotheses. Independence uses no association versus association; homogeneity uses same distributions versus a difference.
Mechanism. Independence uses no association versus association; homogeneity uses same distributions versus a difference. The learner should make the state of the problem visible before using a technique. This turns a remembered rule into a chain of decisions that another reader can inspect. The explanation should identify what information enters the step, what transformation occurs and what result becomes available for the next step.
Worked case. For year group and travel mode, the null says the variables are independent among the school's students. After reaching the result, explain why the step was justified and what evidence would have changed it. That second explanation prevents a correct-looking answer from becoming a lucky routine. It also gives a teacher a precise place to test understanding.
Changed context. Replace one important feature of the case while keeping the underlying decision. Ask what remains stable and what must change. Transfer is demonstrated when the learner preserves the reasoning mechanism without copying surface language, numbers or labels from the first example. A useful comparison names both the invariant decision and the changed evidence.
Failure and repair. Symbols without context or equal groups without a named variable are incomplete. The repair should be used immediately on a near case and then on a farther case. The near case confirms the correction; the farther case checks whether the learner understands the cause rather than the appearance of the original mistake.
Action. Read the hypotheses alone and check whether population and variables remain identifiable. A useful feedback note names the decision, cites the evidence and gives the next small task. It does not merely report that the answer was wrong. Over time, this creates an error history that distinguishes a one-off slip from a recurring misconception.
Student, parent and teacher roles. The student owns the explanation and records the error cause. The teacher chooses the smallest task that can expose the cause and checks a changed case. The parent can protect regular practice, ask the learner to explain the decision aloud and avoid supplying a polished answer that hides the learner's current model.
Evidence boundary. This case is illustrative and original. It demonstrates a reasoning process, not a guaranteed examination response or score. The learner should follow current official instructions and keep each claim within what the prompt, text, design or algorithm can support.
This chapter focuses on calculate expected counts. Under the null, expected count equals row total times column total divided by grand total.
Mechanism. Under the null, expected count equals row total times column total divided by grand total. The learner should make the state of the problem visible before using a technique. This turns a remembered rule into a chain of decisions that another reader can inspect. The explanation should identify what information enters the step, what transformation occurs and what result becomes available for the next step.
Worked case. If 60 of 150 students cycle and 45 are Year 1, expected Year 1 cyclists are 45×60÷150=18. After reaching the result, explain why the step was justified and what evidence would have changed it. That second explanation prevents a correct-looking answer from becoming a lucky routine. It also gives a teacher a precise place to test understanding.
Changed context. Replace one important feature of the case while keeping the underlying decision. Ask what remains stable and what must change. Transfer is demonstrated when the learner preserves the reasoning mechanism without copying surface language, numbers or labels from the first example. A useful comparison names both the invariant decision and the changed evidence.
Failure and repair. Observed-cell averages or wrong margins replace the null model. Label totals first. The repair should be used immediately on a near case and then on a farther case. The near case confirms the correction; the farther case checks whether the learner understands the cause rather than the appearance of the original mistake.
Action. Estimate plausibility, show one calculation and compare with technology output. A useful feedback note names the decision, cites the evidence and gives the next small task. It does not merely report that the answer was wrong. Over time, this creates an error history that distinguishes a one-off slip from a recurring misconception.
Student, parent and teacher roles. The student owns the explanation and records the error cause. The teacher chooses the smallest task that can expose the cause and checks a changed case. The parent can protect regular practice, ask the learner to explain the decision aloud and avoid supplying a polished answer that hides the learner's current model.
Evidence boundary. This case is illustrative and original. It demonstrates a reasoning process, not a guaranteed examination response or score. The learner should follow current official instructions and keep each claim within what the prompt, text, design or algorithm can support.
This chapter focuses on verify conditions first. The effective-fall-2026 framework specifies randomization, the 10% condition when sampling without replacement, and expected counts greater than 5.
Mechanism. The effective-fall-2026 framework specifies randomization, the 10% condition when sampling without replacement, and expected counts greater than 5. The learner should make the state of the problem visible before using a technique. This turns a remembered rule into a chain of decisions that another reader can inspect. The explanation should identify what information enters the step, what transformation occurs and what result becomes available for the next step.
Worked case. A large observed cell cannot rescue a small expected count; an observed zero can still have a sufficient expectation. After reaching the result, explain why the step was justified and what evidence would have changed it. That second explanation prevents a correct-looking answer from becoming a lucky routine. It also gives a teacher a precise place to test understanding.
Changed context. Replace one important feature of the case while keeping the underlying decision. Ask what remains stable and what must change. Transfer is demonstrated when the learner preserves the reasoning mechanism without copying surface language, numbers or labels from the first example. A useful comparison names both the invariant decision and the changed evidence.
Failure and repair. Observed counts are checked, or conditions are recited without evidence. Contextualise each condition. The repair should be used immediately on a near case and then on a farther case. The near case confirms the correction; the farther case checks whether the learner understands the cause rather than the appearance of the original mistake.
Action. Build a condition table that separates design evidence from arithmetic. A useful feedback note names the decision, cites the evidence and gives the next small task. It does not merely report that the answer was wrong. Over time, this creates an error history that distinguishes a one-off slip from a recurring misconception.
Student, parent and teacher roles. The student owns the explanation and records the error cause. The teacher chooses the smallest task that can expose the cause and checks a changed case. The parent can protect regular practice, ask the learner to explain the decision aloud and avoid supplying a polished answer that hides the learner's current model.
Evidence boundary. This case is illustrative and original. It demonstrates a reasoning process, not a guaranteed examination response or score. The learner should follow current official instructions and keep each claim within what the prompt, text, design or algorithm can support.
| Decision | Question | Evidence |
|---|---|---|
| Test family | One population or several populations or treatments? | Sampling and assignment description |
| Expected count | What would the null model predict? | Row total × column total ÷ grand total |
| Contribution | Which cells create discrepancy? | (observed − expected)² ÷ expected |
| Scope | What can the study support? | Random sampling and random assignment |
This chapter focuses on understand cell contributions. Each contribution squares observed minus expected and divides by expected, preventing cancellation and scaling discrepancy.
Mechanism. Each contribution squares observed minus expected and divides by expected, preventing cancellation and scaling discrepancy. The learner should make the state of the problem visible before using a technique. This turns a remembered rule into a chain of decisions that another reader can inspect. The explanation should identify what information enters the step, what transformation occurs and what result becomes available for the next step.
Worked case. Observed 30 versus expected 20 contributes 5; observed 110 versus expected 100 contributes 1. After reaching the result, explain why the step was justified and what evidence would have changed it. That second explanation prevents a correct-looking answer from becoming a lucky routine. It also gives a teacher a precise place to test understanding.
Changed context. Replace one important feature of the case while keeping the underlying decision. Ask what remains stable and what must change. Transfer is demonstrated when the learner preserves the reasoning mechanism without copying surface language, numbers or labels from the first example. A useful comparison names both the invariant decision and the changed evidence.
Failure and repair. Largest raw residual is assumed most important. Compare scaled contributions. The repair should be used immediately on a near case and then on a farther case. The near case confirms the correction; the farther case checks whether the learner understands the cause rather than the appearance of the original mistake.
Action. Rank cells before summing and explain which categories drive the statistic. A useful feedback note names the decision, cites the evidence and gives the next small task. It does not merely report that the answer was wrong. Over time, this creates an error history that distinguishes a one-off slip from a recurring misconception.
Student, parent and teacher roles. The student owns the explanation and records the error cause. The teacher chooses the smallest task that can expose the cause and checks a changed case. The parent can protect regular practice, ask the learner to explain the decision aloud and avoid supplying a polished answer that hides the learner's current model.
Evidence boundary. This case is illustrative and original. It demonstrates a reasoning process, not a guaranteed examination response or score. The learner should follow current official instructions and keep each claim within what the prompt, text, design or algorithm can support.
CHAPTER 7 OF 15 · Use evidence precisely
Compute statistic and degrees of freedom
This chapter focuses on compute statistic and degrees of freedom. Sum contributions; for an r-by-c table degrees of freedom are (r−1)(c−1).
Mechanism. Sum contributions; for an r-by-c table degrees of freedom are (r−1)(c−1). The learner should make the state of the problem visible before using a technique. This turns a remembered rule into a chain of decisions that another reader can inspect. The explanation should identify what information enters the step, what transformation occurs and what result becomes available for the next step.
Worked case. A 3×4 table has six degrees of freedom, reflecting margin constraints rather than sample size. After reaching the result, explain why the step was justified and what evidence would have changed it. That second explanation prevents a correct-looking answer from becoming a lucky routine. It also gives a teacher a precise place to test understanding.
Changed context. Replace one important feature of the case while keeping the underlying decision. Ask what remains stable and what must change. Transfer is demonstrated when the learner preserves the reasoning mechanism without copying surface language, numbers or labels from the first example. A useful comparison names both the invariant decision and the changed evidence.
Failure and repair. Table entry or dimensions are wrong even when a calculator returns an answer. Audit one cell. The repair should be used immediately on a near case and then on a farther case. The near case confirms the correction; the farther case checks whether the learner understands the cause rather than the appearance of the original mistake.
Action. Record labels, hand-check one expectation and verify degrees of freedom. A useful feedback note names the decision, cites the evidence and gives the next small task. It does not merely report that the answer was wrong. Over time, this creates an error history that distinguishes a one-off slip from a recurring misconception.
Student, parent and teacher roles. The student owns the explanation and records the error cause. The teacher chooses the smallest task that can expose the cause and checks a changed case. The parent can protect regular practice, ask the learner to explain the decision aloud and avoid supplying a polished answer that hides the learner's current model.
Evidence boundary. This case is illustrative and original. It demonstrates a reasoning process, not a guaranteed examination response or score. The learner should follow current official instructions and keep each claim within what the prompt, text, design or algorithm can support.
This chapter focuses on interpret the p-value. The p-value assumes the null and measures the probability of a statistic at least as extreme as observed.
Mechanism. The p-value assumes the null and measures the probability of a statistic at least as extreme as observed. The learner should make the state of the problem visible before using a technique. This turns a remembered rule into a chain of decisions that another reader can inspect. The explanation should identify what information enters the step, what transformation occurs and what result becomes available for the next step.
Worked case. A p-value of 0.018 means such aggregate discrepancy would occur about 1.8% of the time under the stated null in comparable sampling. After reaching the result, explain why the step was justified and what evidence would have changed it. That second explanation prevents a correct-looking answer from becoming a lucky routine. It also gives a teacher a precise place to test understanding.
Changed context. Replace one important feature of the case while keeping the underlying decision. Ask what remains stable and what must change. Transfer is demonstrated when the learner preserves the reasoning mechanism without copying surface language, numbers or labels from the first example. A useful comparison names both the invariant decision and the changed evidence.
Failure and repair. It is called the probability the null is true. Restore the conditional statement. The repair should be used immediately on a near case and then on a farther case. The near case confirms the correction; the farther case checks whether the learner understands the cause rather than the appearance of the original mistake.
Action. Use a sentence containing assumption, statistic extremity and context. A useful feedback note names the decision, cites the evidence and gives the next small task. It does not merely report that the answer was wrong. Over time, this creates an error history that distinguishes a one-off slip from a recurring misconception.
Student, parent and teacher roles. The student owns the explanation and records the error cause. The teacher chooses the smallest task that can expose the cause and checks a changed case. The parent can protect regular practice, ask the learner to explain the decision aloud and avoid supplying a polished answer that hides the learner's current model.
Evidence boundary. This case is illustrative and original. It demonstrates a reasoning process, not a guaranteed examination response or score. The learner should follow current official instructions and keep each claim within what the prompt, text, design or algorithm can support.
This chapter focuses on conclude at alpha. Compare p with the significance level, then state evidence for the contextual alternative within scope.
Mechanism. Compare p with the significance level, then state evidence for the contextual alternative within scope. The learner should make the state of the problem visible before using a technique. This turns a remembered rule into a chain of decisions that another reader can inspect. The explanation should identify what information enters the step, what transformation occurs and what result becomes available for the next step.
Worked case. At alpha .05 and p .018, reject no association and report convincing evidence of association. After reaching the result, explain why the step was justified and what evidence would have changed it. That second explanation prevents a correct-looking answer from becoming a lucky routine. It also gives a teacher a precise place to test understanding.
Changed context. Replace one important feature of the case while keeping the underlying decision. Ask what remains stable and what must change. Transfer is demonstrated when the learner preserves the reasoning mechanism without copying surface language, numbers or labels from the first example. A useful comparison names both the invariant decision and the changed evidence.
Failure and repair. A large p is described as proof or acceptance of the null. Say evidence was insufficient. The repair should be used immediately on a near case and then on a farther case. The near case confirms the correction; the farther case checks whether the learner understands the cause rather than the appearance of the original mistake.
Action. Write decision and contextual conclusion as separate linked clauses. A useful feedback note names the decision, cites the evidence and gives the next small task. It does not merely report that the answer was wrong. Over time, this creates an error history that distinguishes a one-off slip from a recurring misconception.
Student, parent and teacher roles. The student owns the explanation and records the error cause. The teacher chooses the smallest task that can expose the cause and checks a changed case. The parent can protect regular practice, ask the learner to explain the decision aloud and avoid supplying a polished answer that hides the learner's current model.
Evidence boundary. This case is illustrative and original. It demonstrates a reasoning process, not a guaranteed examination response or score. The learner should follow current official instructions and keep each claim within what the prompt, text, design or algorithm can support.
This chapter focuses on avoid invented causation. Significance identifies departure from the null. Causation needs random assignment; generalisation needs random sampling.
Mechanism. Significance identifies departure from the null. Causation needs random assignment; generalisation needs random sampling. The learner should make the state of the problem visible before using a technique. This turns a remembered rule into a chain of decisions that another reader can inspect. The explanation should identify what information enters the step, what transformation occurs and what result becomes available for the next step.
Worked case. Association between year and travel mode may generalise from a random sample but advancing a year was not assigned. After reaching the result, explain why the step was justified and what evidence would have changed it. That second explanation prevents a correct-looking answer from becoming a lucky routine. It also gives a teacher a precise place to test understanding.
Changed context. Replace one important feature of the case while keeping the underlying decision. Ask what remains stable and what must change. Transfer is demonstrated when the learner preserves the reasoning mechanism without copying surface language, numbers or labels from the first example. A useful comparison names both the invariant decision and the changed evidence.
Failure and repair. A small p-value is treated as causal proof. Separate pattern, population scope and causal scope. The repair should be used immediately on a near case and then on a farther case. The near case confirms the correction; the farther case checks whether the learner understands the cause rather than the appearance of the original mistake.
Action. Label sampling and assignment before writing the final claim. A useful feedback note names the decision, cites the evidence and gives the next small task. It does not merely report that the answer was wrong. Over time, this creates an error history that distinguishes a one-off slip from a recurring misconception.
Student, parent and teacher roles. The student owns the explanation and records the error cause. The teacher chooses the smallest task that can expose the cause and checks a changed case. The parent can protect regular practice, ask the learner to explain the decision aloud and avoid supplying a polished answer that hides the learner's current model.
Evidence boundary. This case is illustrative and original. It demonstrates a reasoning process, not a guaranteed examination response or score. The learner should follow current official instructions and keep each claim within what the prompt, text, design or algorithm can support.
This chapter focuses on use conditional proportions. Conditional distributions reveal the pattern behind a global test and need denominators aligned with the question.
Mechanism. Conditional distributions reveal the pattern behind a global test and need denominators aligned with the question. The learner should make the state of the problem visible before using a technique. This turns a remembered rule into a chain of decisions that another reader can inspect. The explanation should identify what information enters the step, what transformation occurs and what result becomes available for the next step.
Worked case. Forty percent of Year 1 cycles versus eighteen percent of Year 3; raw counts would mislead if groups differ. After reaching the result, explain why the step was justified and what evidence would have changed it. That second explanation prevents a correct-looking answer from becoming a lucky routine. It also gives a teacher a precise place to test understanding.
Changed context. Replace one important feature of the case while keeping the underlying decision. Ask what remains stable and what must change. Transfer is demonstrated when the learner preserves the reasoning mechanism without copying surface language, numbers or labels from the first example. A useful comparison names both the invariant decision and the changed evidence.
Failure and repair. Percentages use inconsistent denominators. Say out of before every comparison. The repair should be used immediately on a near case and then on a farther case. The near case confirms the correction; the farther case checks whether the learner understands the cause rather than the appearance of the original mistake.
Action. Annotate each proportion with its conditioning group. A useful feedback note names the decision, cites the evidence and gives the next small task. It does not merely report that the answer was wrong. Over time, this creates an error history that distinguishes a one-off slip from a recurring misconception.
Student, parent and teacher roles. The student owns the explanation and records the error cause. The teacher chooses the smallest task that can expose the cause and checks a changed case. The parent can protect regular practice, ask the learner to explain the decision aloud and avoid supplying a polished answer that hides the learner's current model.
Evidence boundary. This case is illustrative and original. It demonstrates a reasoning process, not a guaranteed examination response or score. The learner should follow current official instructions and keep each claim within what the prompt, text, design or algorithm can support.
This chapter focuses on diagnose sparse cells. Low expected counts may reflect limited data or overfine categories; combining is defensible only with substantive meaning.
Mechanism. Low expected counts may reflect limited data or overfine categories; combining is defensible only with substantive meaning. The learner should make the state of the problem visible before using a technique. This turns a remembered rule into a chain of decisions that another reader can inspect. The explanation should identify what information enters the step, what transformation occurs and what result becomes available for the next step.
Worked case. Bus and train may combine for public transport policy but not for a reliability comparison. After reaching the result, explain why the step was justified and what evidence would have changed it. That second explanation prevents a correct-looking answer from becoming a lucky routine. It also gives a teacher a precise place to test understanding.
Changed context. Replace one important feature of the case while keeping the underlying decision. Ask what remains stable and what must change. Transfer is demonstrated when the learner preserves the reasoning mechanism without copying surface language, numbers or labels from the first example. A useful comparison names both the invariant decision and the changed evidence.
Failure and repair. Categories are merged after seeing results merely to force conditions. Return to the question. The repair should be used immediately on a near case and then on a farther case. The near case confirms the correction; the farther case checks whether the learner understands the cause rather than the appearance of the original mistake.
Action. Document grouping before inference or seek more data or another method. A useful feedback note names the decision, cites the evidence and gives the next small task. It does not merely report that the answer was wrong. Over time, this creates an error history that distinguishes a one-off slip from a recurring misconception.
Student, parent and teacher roles. The student owns the explanation and records the error cause. The teacher chooses the smallest task that can expose the cause and checks a changed case. The parent can protect regular practice, ask the learner to explain the decision aloud and avoid supplying a polished answer that hides the learner's current model.
Evidence boundary. This case is illustrative and original. It demonstrates a reasoning process, not a guaranteed examination response or score. The learner should follow current official instructions and keep each claim within what the prompt, text, design or algorithm can support.
This chapter focuses on transfer to changed designs. The reasoning chain persists when table labels and dimensions change, but scope follows design.
Mechanism. The reasoning chain persists when table labels and dimensions change, but scope follows design. The learner should make the state of the problem visible before using a technique. This turns a remembered rule into a chain of decisions that another reader can inspect. The explanation should identify what information enters the step, what transformation occurs and what result becomes available for the next step.
Worked case. A randomized treatment-by-symptom table may support causation while population generalisation still depends on recruitment. After reaching the result, explain why the step was justified and what evidence would have changed it. That second explanation prevents a correct-looking answer from becoming a lucky routine. It also gives a teacher a precise place to test understanding.
Changed context. Replace one important feature of the case while keeping the underlying decision. Ask what remains stable and what must change. Transfer is demonstrated when the learner preserves the reasoning mechanism without copying surface language, numbers or labels from the first example. A useful comparison names both the invariant decision and the changed evidence.
Failure and repair. Memorised wording from a survey is pasted into an experiment. Rewrite every contextual noun. The repair should be used immediately on a near case and then on a farther case. The near case confirms the correction; the farther case checks whether the learner understands the cause rather than the appearance of the original mistake.
Action. Practise mixed scenarios without the procedure name supplied. A useful feedback note names the decision, cites the evidence and gives the next small task. It does not merely report that the answer was wrong. Over time, this creates an error history that distinguishes a one-off slip from a recurring misconception.
Student, parent and teacher roles. The student owns the explanation and records the error cause. The teacher chooses the smallest task that can expose the cause and checks a changed case. The parent can protect regular practice, ask the learner to explain the decision aloud and avoid supplying a polished answer that hides the learner's current model.
Evidence boundary. This case is illustrative and original. It demonstrates a reasoning process, not a guaranteed examination response or score. The learner should follow current official instructions and keep each claim within what the prompt, text, design or algorithm can support.
This chapter focuses on diagnose reasoning errors. A wrong number can arise from entry, expectation or degrees of freedom; a correct number can coexist with wrong scope.
Mechanism. A wrong number can arise from entry, expectation or degrees of freedom; a correct number can coexist with wrong scope. The learner should make the state of the problem visible before using a technique. This turns a remembered rule into a chain of decisions that another reader can inspect. The explanation should identify what information enters the step, what transformation occurs and what result becomes available for the next step.
Worked case. Correct calculator output followed by a causal claim from a survey requires design practice, not more arithmetic. After reaching the result, explain why the step was justified and what evidence would have changed it. That second explanation prevents a correct-looking answer from becoming a lucky routine. It also gives a teacher a precise place to test understanding.
Changed context. Replace one important feature of the case while keeping the underlying decision. Ask what remains stable and what must change. Transfer is demonstrated when the learner preserves the reasoning mechanism without copying surface language, numbers or labels from the first example. A useful comparison names both the invariant decision and the changed evidence.
Failure and repair. Whole-test repetition hides the broken link. Tag each response step. The repair should be used immediately on a near case and then on a farther case. The near case confirms the correction; the farther case checks whether the learner understands the cause rather than the appearance of the original mistake.
Action. Keep a reasoning ledger and practise the smallest failed decision. A useful feedback note names the decision, cites the evidence and gives the next small task. It does not merely report that the answer was wrong. Over time, this creates an error history that distinguishes a one-off slip from a recurring misconception.
Student, parent and teacher roles. The student owns the explanation and records the error cause. The teacher chooses the smallest task that can expose the cause and checks a changed case. The parent can protect regular practice, ask the learner to explain the decision aloud and avoid supplying a polished answer that hides the learner's current model.
Evidence boundary. This case is illustrative and original. It demonstrates a reasoning process, not a guaranteed examination response or score. The learner should follow current official instructions and keep each claim within what the prompt, text, design or algorithm can support.
This chapter focuses on keep the 2027 boundary current. The revised CED retains homogeneity and independence with explicit conditions, while goodness-of-fit was removed; May 2027 is fully digital.
Mechanism. The revised CED retains homogeneity and independence with explicit conditions, while goodness-of-fit was removed; May 2027 is fully digital. The learner should make the state of the problem visible before using a technique. This turns a remembered rule into a chain of decisions that another reader can inspect. The explanation should identify what information enters the step, what transformation occurs and what result becomes available for the next step.
Worked case. Older examples may still teach arithmetic but can misstate current scope and delivery. After reaching the result, explain why the step was justified and what evidence would have changed it. That second explanation prevents a correct-looking answer from becoming a lucky routine. It also gives a teacher a precise place to test understanding.
Changed context. Replace one important feature of the case while keeping the underlying decision. Ask what remains stable and what must change. Transfer is demonstrated when the learner preserves the reasoning mechanism without copying surface language, numbers or labels from the first example. A useful comparison names both the invariant decision and the changed evidence.
Failure and repair. Materials are used without checking their effective year. Match every resource to the current CED. The repair should be used immediately on a near case and then on a farther case. The near case confirms the correction; the farther case checks whether the learner understands the cause rather than the appearance of the original mistake.
Action. Date notes, use original tables and avoid score promises. A useful feedback note names the decision, cites the evidence and gives the next small task. It does not merely report that the answer was wrong. Over time, this creates an error history that distinguishes a one-off slip from a recurring misconception.
Student, parent and teacher roles. The student owns the explanation and records the error cause. The teacher chooses the smallest task that can expose the cause and checks a changed case. The parent can protect regular practice, ask the learner to explain the decision aloud and avoid supplying a polished answer that hides the learner's current model.
Evidence boundary. This case is illustrative and original. It demonstrates a reasoning process, not a guaranteed examination response or score. The learner should follow current official instructions and keep each claim within what the prompt, text, design or algorithm can support.
Sources: AP Statistics CED effective fall 2026 · AP Statistics exam information · AP Statistics revisions
Related reading: Education parent guide · AQA Psychology paired-data test choice · IB Mathematics explorations and numerical bounds