Score Scale and Testlets

Why the CPA Score Scale Matters

A passing CPA Exam score is 75, but that number is widely misunderstood. Scores are reported on a 0-99 scale, and 75 is the fixed passing standard for every section: Auditing and Attestation (AUD), Financial Accounting and Reporting (FAR), Regulation (REG), Business Analysis and Reporting (BAR), Information Systems and Controls (ISC), and Tax Compliance and Planning (TCP). The 75 is a scaled score, not 75 percent correct, and not a percentile rank. A candidate who sees 74 did not necessarily miss one more question than one who scored 75.

Scaled, Not Curved

The AICPA states scores are not curved. Your result is never lowered because another testing group performed well, nor raised because a cohort struggled. The exam is assembled from different but psychometrically comparable forms, and scaling ensures a 75 means the same thing across forms and over time. This is why a plan built on raw counts, such as "I need 55 of 72 REG MCQs," is unreliable: each item carries a different difficulty weight, so a harder question correctly answered contributes more than an easy one.

Pretest Items Hide in Plain Sight

Every section contains pretest multiple-choice questions (MCQs) and task-based simulations (TBSs) that gather statistical data for future exams. They are indistinguishable from operational, scored items, and they do not count toward your reported score. Because you cannot tell which is which, the only rational behavior is to treat every item as scored and refuse to panic over one strange prompt — it may simply be an experimental item being calibrated.

Five Testlets, One Direction

Each section is a four-hour exam built from five testlets. Testlets 1 and 2 are MCQs; testlets 3, 4, and 5 are TBSs. Inside the current testlet you may move freely, change answers, and review flagged items. Once you submit, that testlet locks — you cannot revisit an MCQ after seeing later simulations, and you cannot advance without submitting.

What This Means While Studying

Make the testlet structure automatic. Practice MCQs in timed sets so you decide without rereading a stem four times. Practice simulations in sets too, because the last three testlets demand sustained document reading, calculation, and response entry after you have already burned mental energy on roughly 70-82 MCQs.

A disciplined in-testlet routine:

• Flag only with a real reason to revisit — a coin-flip between two options, or a number you want to recompute.
• Answer every item first, then circle back to flags; never leave a blank because the scaled score gives zero credit for it.
• Do not waste seconds guessing whether an item is pretest — that guess earns nothing.
• Build the muscle of submitting confidently once review is done, because finality is baked into the exam.

Common trap: treating 75 as "answer three-quarters correctly." A candidate can answer fewer than 75 percent of items yet still score 75 if their correct items skew difficult, and vice versa. Aim for mastery and consistent accuracy, not a raw fraction.

How Scaling Actually Works

The psychometric model behind the CPA Exam is item response theory (IRT). Under IRT, every operational item has a calibrated difficulty parameter, and your scaled score is computed from which items you answered correctly, not merely how many. Two candidates can each answer 60 of 80 items correctly and finish with different reported scores if one cleared harder items. This is why your review software's percentage is only a rough proxy: it weights every question equally, but the live exam does not.

MCQs are also delivered multistage adaptively across the two MCQ testlets. Testlet 1 is medium difficulty. If you perform well, testlet 2 is drawn from a harder pool; if you struggle, testlet 2 is moderate. Crucially, a harder testlet 2 is good news — it signals strong testlet-1 performance, and its tougher items carry more scaled value, so a 'medium-medium' path and a 'medium-hard' path can both pass. Do not panic if the second MCQ testlet feels brutal; that difficulty is partly a reward, and the scaling already accounts for it.

TBS testlets, by contrast, are not adaptive — every candidate sees a comparable simulation pool, and your three simulation testlets do not change based on earlier performance.

Score Release and Reading Your Report

Scores post on the AICPA's published target release dates after the close of a quarterly scoring window, and they appear on your NASBA candidate portal as a single number from 0-99. A failing report includes diagnostic feedback comparing your performance to passing candidates by content area and item type (Stronger, Comparable, Weaker). Use that feedback to retarget study; it tells you whether MCQ accuracy or simulation execution sank the attempt, which directly informs the weighting decisions in the next section.