How the 2026 ExAC Is Marked: The Angoff Method Explained

The ExAC has no fixed pass percentage. Each section's passing score comes from the Angoff method: architects trained as markers estimate how a minimally competent candidate would do on every question, and their estimates are averaged into a pass line. Our team went through ExAC's own reports to explain how it works. Examitect is an independent prep platform, unaffiliated with the CACB.

Key Takeaways

What the Angoff method means for you.

  • Angoff sets the pass line. It does not mark your paper. An optical reader marks multiple choice, and trained architects mark the written answers in Section 1 and Section 3.
  • ExAC names the method itself. Its passing score page lists "the Angoff method and its variants", and its technical report cites Angoff's 1971 work.
  • No pass percentage is published. Passing scores are not made public, and your transcript shows only pass or fail for each section.
  • Each section has its own passing score. A strong Section 4 cannot rescue a weak Section 2.
  • Close calls get the benefit of the doubt. A margin of error sits below the panel's estimate, and a fail means your score landed below that margin as well.
  • Study for breadth and answer everything. A blank multiple-choice answer counts as incorrect, and no practice score converts to an ExAC result, so use mock exams for pacing and review.

Overview

ExAC marking at a glance

MethodAngoff and its variants, with statistical checks
Who rates the questionsArchitects trained as ExAC markers
Last full exerciseFall 2024, after Theme 13 was added
When the line is setAfter each sitting, once results are analysed
Published pass markNone; passing scores are not made public
What you receivePass or fail per section, no raw score
2026 resultsAround mid-February, by mail from your licensing authority

What is the Angoff method?

The Angoff method sets a pass line by judging the questions, not by ranking the people who wrote them. A panel of subject experts pictures a candidate who is just good enough to enter the profession. For each question, each panellist estimates the chance that this candidate gets it right. Average those estimates and you have the score that candidate should earn, which becomes the basis of the pass line.

The method takes its name from William Angoff, whose 1971 chapter in Educational Measurement is the work ExAC's own technical report cites. It is widely used on licensing and certification exams. Canada's physician assistant certification exam, for one, uses a modified form of it.

The minimally competent candidate is the idea that makes it work. This is neither the average candidate nor a star. It is the person who has reached the minimum level of competence needed to practise, and no more. Panellists are not asking whether a question is hard in general. They are asking how this one person would do on it.

A worked example

Here is the arithmetic on a five-question exam with three panellists. The questions and numbers are invented to show the method. They do not come from any ExAC paper or panel.

Question typePanellists A, B, CAverage
Recall of a definition80%, 85%, 75%80%
One-step code lookup65%, 60%, 70%65%
Envelope detail with a flaw to spot70%, 60%, 65%65%
Contract scenario, two plausible answers45%, 55%, 50%50%
Multi-step cost calculation40%, 50%, 45%45%
Panel's estimate for this paperAverage of the five61%

Now suppose next year's paper is harder, and the panel rates the same five slots at 70, 55, 55, 40 and 35 percent. The estimate drops to 51 percent. That is the point of the method: a harder paper gets a lower line, an easier paper gets a higher one, and the standard expected of the candidate stays where it was.

How does the ExAC use the Angoff method?

ExAC is open about the method but not about the numbers. Its passing score page names Angoff, and the fuller account is in the 2025 Statistical and Technical Report (pages 33 to 35), the ExAC Framework Document (page 15) and the French edition of the report, which gives a little more detail than the English. Together they describe five steps.

  1. Architects rate every question. The panellists are architects who were trained as markers at an earlier scoring session. The French report says they rate each item individually after training, then debate the items where their judgments differ significantly.
  2. The ratings are averaged. The average is the panel's estimate of the minimum score a minimally competent candidate should earn on each section.
  3. Repeated questions keep the years comparable. Each exam repeats 10 to 20 percent of the previous year's paper. Scores on those common items are tracked through a process called equating, so a candidate is not helped or hurt by writing in an easier or harder year.
  4. A margin of error is applied. No exam measures perfectly, so a confidence interval is calculated around the panel's estimate. The Framework calls this the zone of confidence.
  5. Specialists check it and the committee approves. Statistical test analysis experts verify that the passing scores are valid and reliable, and the results go to the Committee for the ExAC (CExAC) for approval.

Step four matters most to anyone waiting on a result. Your score is not compared with a single hard line.

Where your score landsResult
At or above the panel's estimatePass
Below the estimate, but inside the margin of errorPass, with the benefit of the doubt
Below the lower limit of the marginFail

The panel's average is also not the only input. The summary of the 2025 report calls it one of the elements considered, and the full report lists contextual and historical references and the changing difficulty of common items alongside it. The line is set after the sitting, once results have been analysed, so nobody in the exam room knows exactly where it will fall.

Has the standard changed for 2026?

The most recent standard setting on record took place in fall 2024. The Framework ties it to the addition of Theme 13, Sustainable Design Literacy. According to the 2025 report, the exercise runs in whole or in part every few years, or after a major event such as COVID-19 or a substantial change in exam content. A new edition of the National Building Code is the example it gives.

The 2024 exercise moved things only slightly. The report summarizes it this way (page 35):

What was measured2024 result
Difficulty of common items, as monitoredUp 1.57 percentage points
Performance of minimally competent architectsPassing rate down 2.1 percentage points
Overall decisionPassing rate lowered by about 1.5 percentage points

The report reads this as a sign that current candidates are performing somewhat lower. It uses the words passing rate without saying whether that means the pass mark or the share of candidates who pass, and it publishes neither figure. Read it for direction and size: a small adjustment, downward.

As of early November 2026, nothing ExAC has published says whether a new panel rated this year's paper. What we do know is that each section's passing score is fixed after the sitting. The report on the 2026 exam should follow within a few months. ExAC posts its current report on its development page.

What percentage do you need to pass the ExAC?

Nobody outside the process can tell you, because ExAC does not make its passing scores public. A firm number from any prep site, ours included, would be a guess. The official documents also describe no fixed percentage, since each passing score depends on the content and difficulty of that exam's questions.

The nearest thing to a public figure is an older one. The 2023 technical report (page 30) observed that the cut-off point had tended to sit around 50 percent of the points on the items retained. That sentence does not appear in the 2024 report or the 2025 report, and the standard was revised in between. The 2025 report's figure still draws the lower limit near 50 on a scale of 0 to 100, but it presents that figure as an illustration of the factors involved, not as a pass mark.

Please do not turn 50 percent into your target. Here is why:

  • It is dated and approximate. The observation predates the current standard, and ExAC has not repeated it.
  • It does not describe your mock exam. It counts points on the items kept after analysis, on a paper built to separate stronger candidates from weaker ones. Practice exams, ours included, use different questions, so no practice percentage converts to an ExAC result. We would rather tell you that than sell you a predictor.
  • Aiming at the line leaves you no margin. You have to clear it four times, once per section, in three hours each.

What ExAC does publish is how many candidates pass.

SittingPassed all sections they were eligible to writeAverage pass rate per section
202373.6%85.2%
202472.2%83.5%
202573.8%84.7%

Those rates stayed within a couple of points across the 2024 revision. None of these documents describes a quota or a curve. Score distribution was among the factors reviewed in 2024, so candidate results do inform the process, but nothing ExAC publishes has you competing for a limited number of passes. That should take some weight off.

How are your answers marked?

Angoff sets the line. Marking produces the score measured against it, and this is the part you can influence on exam day.

Multiple-choice answers go through an optical mark reader, and every section has them. Written answers appear only in Sections 1 and 3, and they are marked by trained architects using scoring guides and sample answers chosen to show the range of performance. At least two markers score each answer, and neither can see the other's score. If they agree exactly, that score stands. If they do not, a third marker scores the answer and the average is awarded.

Not every question ends up counting. New questions are field-tested inside the real exam and look identical to the others, and those that perform poorly stay out of your score. Analysis after the sitting can also exclude items found to be inadequate. You cannot tell which is which, so treat every question as live.

The 2026 ExAC Preparation Guide (pages 7 to 9) sets out how answers are read. Each rule below is easy to follow once you know it.

The Prep Guide's ruleWhat to do
A blank multiple-choice answer counts as incorrectBubble an answer for every question
If a question asks for two points, only the first two are readGive the number asked for, clearly separated
Answers outside the lines or on scratch paper are not markedWrite concisely and stay in the space provided
Points are awarded for each calculation stepShow every step, even simple addition
Markers score what you wrote and do not interpret itState your answer plainly and legibly
Spelling and grammar do not count against youSpend your time on content
Questions worth more points are clearly identifiedGive those questions priority

What this means for how you study

ExAC does not give study advice based on its standard setting, so what follows is our own reading, shaped by working with Intern Architects through their preparation.

Cover every theme before you polish any. The candidate the panel pictures handles routine questions and misses plenty of hard ones. So a pass tends to come from collecting the marks a competent architect is expected to earn across the whole paper, and a theme you skipped gives those marks away. The Prep Guide lists the objectives under all thirteen themes, and our ExAC structure guide explains how the sections are built.

Treat each section as its own exam. Each one has its own passing score, and surplus marks do not carry over. Put your extra hours on your weakest section, using our guides to Section 1, Section 2, Section 3 and Section 4.

Read practice scores as a trend. Compare a score with your own last attempt, not with a rumoured pass mark, and use your wrong answers to find the objectives to revisit. Start practice questions early. Our 12-week ExAC study plan builds to 500 to 800 timed questions across a cycle.

Collect the marks the rules give you. Answer every multiple-choice question, show your calculation steps, and do not stall on one hard question when the Prep Guide tells you to move on and come back. Our exam day guide has a four-step pacing plan for a mixed section.

After the 2026 sitting: results, rewrites and reviews

The 2026 ExAC was written on November 2 and 3. Your licensing authority mails results around mid-February as a transcript marked pass or fail for each section. Raw scores are not included, so you will not learn how close you came either way. If you failed a section, the report gives constructive information on the parts that went less well.

A failed section is a setback you can plan around. You get three consecutive attempts, and sections you passed stay passed. Register for the next sitting, because skipping it counts as a failed attempt. If sections remain after a third attempt, you rewrite all four as a first-time candidate. Our retake plan walks through the next steps.

You can also request a review for $300, however many sections are involved. Multiple-choice sheets are checked by hand against the answer key, and short-answer booklets are marked again by trained architects. The fee is refunded if a fail becomes a pass. You receive only a pass or fail, with no number, and the outcome is final. Your form and payment go to your licensing authority before its deadline.

Keep in mind what a review can do. A fail has already landed below the margin of error, so a review looks for a marking or scanning error. It does not reopen where the line was set. Whichever way February goes, a pass line built around minimum competence is a fair thing to prepare for, and most candidates clear it section by section.

FAQ

Angoff method and ExAC marking FAQ

The Angoff method is how the ExAC sets its passing scores. Architects trained as markers estimate the chance that a minimally competent candidate would answer each question correctly, and their estimates are averaged. ExAC also uses repeated questions and statistical analysis to check that the passing score for each section is valid and comparable from year to year.

ExAC does not publish a pass percentage, and its passing scores are not made public. They depend on the content and difficulty of each exam. The 2023 technical report observed that the cut-off point had tended to sit around 50 percent of the points on retained items, but the standard was revised in 2024, so treat that as history and not as a target.

The ExAC is not described as curved in any document it publishes, and none mentions a quota of passes. Passing scores are based on the content and difficulty of the questions and on what a minimally competent candidate should achieve. Score distribution was one of the factors reviewed in the 2024 standard setting, so candidate results do inform the process.

According to the 2025 technical report, the ExAC runs its standard-setting process in whole or in part every few years, or after a major event such as a substantial change in exam content. It was last performed in 2024. Passing scores are still established after each sitting, and repeated questions are tracked to keep the years comparable.

Each of the four ExAC sections has its own passing score and is reported as pass or fail. A high result on one section does not offset a low result on another. Sections you pass stay passed during your three consecutive attempts, so you rewrite only the ones you failed.

Trained architects mark the ExAC short answers, which appear in Sections 1 and 3. At least two markers score each answer without seeing each other's scores. If the two scores do not match exactly, a third marker scores the answer and the average is awarded. An optical mark reader handles the multiple-choice answers.

You cannot get a raw ExAC score. The transcript mailed around mid-February shows pass or fail for each section, and a failed section comes with constructive information on what went less well. A request for review also returns only a pass or fail, with no number attached.

An ExAC review costs $300 regardless of the number of sections, and the fee is refunded if a fail becomes a pass. It checks multiple-choice sheets by hand and marks short answers again, so it looks for marking errors. A fail has already landed below the margin of error, so we suggest putting your main energy into your rewrite plan.