Chi-square goodness of fit and slope inference are out

College Board has revised AP Statistics. The new course takes effect in the 2026-27 school year, and the May 2027 exam is the first one built from it. If you start the course this fall, this is your exam.

AP Central's revisions page names what was removed, by topic number. The chi-square goodness of fit test, topics 8.2 and 8.3 in the old outline, is gone. So is all of Unit 9, inference for slopes.

In plain terms, goodness of fit tested whether one categorical variable followed a claimed distribution. The classic classroom example is checking whether a die is fair. Inference for slopes covered confidence intervals and significance tests for the slope of a regression line, and it filled the last unit of the old course.

One boundary is easy to miss. The chi-square family did not disappear. The tests for homogeneity and for independence, which compare groups or check whether two variables are related, stay in the course. Only goodness of fit was cut.

Three quieter removals ride along

The headline was two topics, but the same list drops three more. Analyzing departures from linearity, topic 2.9, covered transforming curved data to straighten it before fitting a line. Combining random variables, topic 4.9, held the rules for the mean and variance of sums and differences. The geometric distribution, topic 4.12, modeled the number of trials until a first success.

None of these three made the announcement's headline, and that is exactly the problem. All of them show up constantly in older practice sets, usually as quick multiple-choice questions. Materials written for the old course will keep asking about them as if they still count.

Nine units fold into five, and the algebra gate drops

The structure changed along with the content. Units 1 through 8 were consolidated into five units, and topics were reordered inside them. The old list of course skills was replaced by four statistical practices: formulate questions, collect data, analyze data, and interpret results. New material on writing investigative questions was added.

The entry requirement moved too. The old prerequisite of second-year algebra was dropped, and a first-year algebra course is now enough. College Board describes the revision as bringing AP Statistics in line with current introductory college statistics courses.

For students, the practical piece is the topic list, not the unit numbers. The surviving content now lives at different addresses. A video series keyed to the old Unit 7 may cover material that sits in a different unit now, so match by topic name, never by unit number.

May 2027 is fully digital, with a different shape

The exam changes go past the topic list. The multiple-choice section grows from 40 to 42 questions, and each question drops from five answer choices to four. Two new sets of three linked questions appear, one on probability and random variables, one on regression.

The free-response section shrinks from six questions to four, and each question grows from 4 points to 10. Every response is typed and submitted through Bluebook, College Board's testing app. Paper free-response booklets are gone.

Fewer, heavier questions change the risk. On the old exam, one bad free response was one question of six. On the new exam it is one of four, and each one carries a quarter of the section. Practicing complete, written-out responses matters more than it ever did.

What your class covers and what the exam tests are separate lists

A syllabus is a school's decision. The exam's topic list belongs to College Board. Those two can disagree, and 2026-27 is a year when they will, in many classrooms.

Your teacher may still teach goodness of fit or slope inference. Some will, because college courses use them and because units built over years do not vanish on command. Learning them is not wasted education. It simply earns no points in May 2027.

The reverse gap is bigger. Nearly every AP Statistics prep book, video series, and question set in circulation was written for the old course. Those materials will keep drilling the five removed topics as if they were required. The questions are not wrong. They test a course you are no longer taking. The checking falls to you and the revised course description.

Keep the revised topic list open next to every practice set

Download the revised course description from AP Central and treat it as the referee for study materials. When an older practice question covers goodness of fit, slope inference, geometric distributions, combining random variables, or straightening curved data, you are looking at removed content. It cannot earn points on the May 2027 exam.

That is a fact about the exam, not an instruction to skip schoolwork. If your teacher still assigns those topics, do the work. Your course has its own requirements, its own grade, and reasons of its own, including the college statistics courses ahead. The change governs a narrower decision: which materials you pick for your own exam review. Match those to the revised topic list, not the old one.

Point your own review hours where old materials cannot help: the two new linked multiple-choice sets and the four 10-point free-response questions. May 2027 is the first exam in this format, so no authentic past exam exists for it yet. College Board's own practice resources for the revised course are the ones to watch as they appear.

And in class, when your teacher covers a topic you cannot find in the revised course description, ask one question early: is this on the exam, or is this for later. Both answers are useful, and neither is a reason to tune out. Only one of them should change which practice materials you reach for.