AP Stats reference sheet: What's given, what to memorize

The AP (Advanced Placement) Statistics reference sheet (officially, the formula sheet and tables) is provided in both sections of the exam, in print and inside the Bluebook testing app. The course was revised for 2026-27, so check the formula sheet in the appendix of the current Course and Exam Description to make sure you practice with the latest version. It covers descriptive stats, sampling distributions, every confidence interval and significance test, and three probability tables. But it doesn't cover everything, and students who assume it does spend exam time hunting for formulas that were never printed.

This guide walks through what the sheet gives you, what it deliberately leaves off, and what you still need to have in memory on exam day. If you'd rather see every formula explained with a worked example, Cognito's AP Statistics past papers work through recent released exams by year.

The sheet is generous, but it deliberately omits definitions, rules and interpretations. You still need these cold.

Conditions for inference procedures. The formulas are printed; the conditions to check are not. For every t-procedure you need to remember random sample, 10 percent condition (n less than 10 percent of population), and Normal or large sample. For proportions you need random sample, 10 percent condition, and np and n(1-p) at least 10. Missing a condition check on an FRQ costs the conditions rubric point.

Interpretation wording. Confidence interval interpretation ('we are C percent confident that the true parameter lies between L and U') is a rubric point on almost every FRQ that involves an interval. The sheet doesn't remind you of the wording; you need it memorized.

Type I and Type II errors, and power. Definitions, relationships and consequences aren't on the sheet. Recent FRQs have asked students to identify a Type I error in context and describe a consequence, and the rubric wants the specific language.

Probability rules. The sheet prints the general addition rule, P(A or B) = P(A) + P(B) - P(A and B), and the conditional probability formula, P(A given B) = P(A and B) / P(B), but not when to use them: you still need to know when events are mutually exclusive or independent, and how to rearrange the conditional formula into the multiplication rule.

Experimental design vocabulary. Blocking, stratification, random assignment, placebo, blinding, confounding, lurking variable. All fair game on FRQ 2 or 3 in a typical year. None printed on the sheet.

The last pages of the reference sheet are three probability tables. You can use them directly, or use the calculator equivalent; both are accepted on the exam.

Standard normal table (Table A). Gives P(Z less than z) for z from about minus 3.4 to 3.4 in steps of 0.01. Look up the row for the first two digits of z, the column for the third. Calculator equivalent on the TI-84: normalcdf(lower, upper), or normalcdf(-1E99, z) for P(Z less than z).

t-distribution table (Table B). Rows are degrees of freedom, columns are one-tail alpha levels and confidence levels. Use it to find critical t* for a confidence interval or to bracket a p-value. Calculator equivalent: tcdf(lower, upper, df) for p-values, invT(area, df) for critical values.

Chi-square table (Table C). Rows are df, columns are common alpha levels. Same use as the t table: bracket a p-value or find a critical value. Calculator equivalent: chisquaredcdf(chi-squared, 1E99, df) for right-tail p-value; the TI-84 has no inverse chi-square command, so read critical values from Table C.

Best practice: sit every practice FRQ with the sheet open on your desk from the very first practice paper. That builds the muscle of scanning the sheet efficiently rather than reading it word-by-word under exam pressure.

For each FRQ, before you touch a formula, identify the procedure: 'this is a two-sample t-test because we have two independent samples of a quantitative variable and we're testing whether the means differ'. Then reach for the formula. Doing it in the other order (formula first, then justification) is how students fit the wrong procedure to the data.

On the MCQ section, the sheet is useful for direct-computation questions. For interpretation MCQs (which are more common in recent years) it's less directly helpful, but the tables at the back are useful for checking critical values quickly if your memory of the standard 1.96 for 95 percent slips.

On the MCQ section, the sheet doesn't save you from the interpretation-style questions that make up a large share of recent papers. When a question shows you a boxplot and asks 'which conclusion is best supported', no formula on the sheet helps; the answer depends on knowing what a boxplot can and can't reveal (it shows median and IQR clearly but can hide bimodality). Sheet-only revision misses this.

On the FRQ section, the sheet doesn't save you from the four-step inference writing structure the rubric expects. State the hypotheses in symbols and words, plan the procedure with conditions checked in context, do the computation with the calculator or by hand, and conclude in context with a specific reference to the parameter under test. Rubrics award points for each of the four steps separately, so an answer that jumps straight to the calculator output loses points even when the numerical answer is right.

Revision priority: content mastery first (know when each procedure applies), then procedural fluency (know the four-step template cold), then reference-sheet familiarity (know where each formula sits so you can find it in five seconds). Skipping the first two in favor of just memorizing the sheet is a common mistake.