Do Certification Exams Give Partial Credit? How Multiple-Response Questions Are Actually Scored

Table of Contents
Executive Summary
The three big vendors answer this differently, and the difference is not cosmetic. A per-domain practice diagnostic is the only way to see which of these scoring models you are actually exposed to. PMI scores each scored question as one point, all or nothing, then converts the total into a performance band. ISC2's exams are compensatory — a strong domain can offset a weak one — and you get a single pass/fail with no number. CompTIA does not disclose its scoring method, converts your result to a scaled score, and refuses to say how many questions you answered incorrectly.
- PMI: one point per scored question. No partial credit within an item, but the domains are weighted, so a weak domain can still be compensated by weight.
- ISC2: explicitly compensatory across domains, pass/fail, no numerical score provided.
- CompTIA: a mathematical algorithm producing a scaled score. Passing scores are set by statistical analysis and can change.
A multiple-response question with two correct answers out of five is a different bet depending on who wrote the exam. Treating "roughly half right" as partial credit is the most common miscalculation in certification preparation, and it is wrong in a specific way for each vendor.
PMI: one point per question, then a band, not a number
PMI's published score report example is unambiguous: "Each scored question on the exam is worth one point; and your final score is calculated by totaling the points you have earned on the exam." A multiple-response question is worth one point, so it is all or nothing — picking two of three correct answers scores the same as picking none.
What the report does not contain is a number you can check yourself against. PMI states that "the number of questions you answer correctly places you within one of the performance rating categories" and the categories are:
- Above Target — performance exceeds the minimum requirements.
- Target — performance meets the minimum requirements.
- Below Target — slightly below target, does not meet the minimum.
- Needs Improvement — far below target, does not meet the minimum.
These are calculated within each domain, and the report includes a "Your Performance by Domain" section. The cut is set by subject matter experts through "sound psychometric analysis", and PMI describes the exams as criteria-based. Any percentage you have read for the PMP — 60%, 65% — is somebody's estimate, not PMI's.
One consequence worth internalising: the score at the end of your session is preliminary, and the official report can take up to 10 business days to arrive.
ISC2: compensatory by design, and no score at all
ISC2's CISSP documentation says the exam is "a single pass/fail result calculated on the total of all operational items" and, more importantly, that "ISC2 exams are compensatory exams which allow for a higher number of items answered correctly in one domain to compensate for a lower performance in another domain."
That is a materially different model from PMI's. In a compensatory exam, a weak domain is not automatically fatal, because the domains are not weighted independently — the total operational set carries the result.
What you get is even less than a score. ISC2's own page answers the question directly: "No scores are provided." Failing candidates do receive a rank ordering of domains by level — below proficiency, near proficiency, above proficiency — at the testing center, and that is it. Candidates are not told how many questions they got right or what percentage that was.
Note that the pre-exam guidance mentions a scale score of at least 700 out of 1,000 to pass. That is the standard for the ISC2 cybersecurity exams, and the exams themselves are scored pass/fail on the operational items — a distinction that catches people out.
CompTIA: the method is deliberately not public
CompTIA's position is that it will not tell you how it works, and says so plainly: "CompTIA does not release specific exam questions, answers, scoring keys, or detailed scoring information."
What it does publish is the shape of the output. A mathematical algorithm produces a "scaled score" that is "derived from what would otherwise be a percentage score", and passing scores are "set using statistical analysis and are subject to change". Security+ V7 uses a 100-900 scale with a passing score of 750.
Two consequences. First, your score report lists "a list of exam objectives associated with questions that were answered incorrectly" — which tells you which objectives to revise, but not which questions you got wrong or how many. Second, a result can come back indeterminate, in which case "CompTIA will allow candidates whose scores are determined to be indeterminate the option of retaking the exam at no cost" — but an indeterminate score "will not be considered valid and will not be eligible for certification".
CompTIA also does not publish pass rates, on the grounds that questions and pass rates change without notice. If you have seen a CompTIA pass rate quoted, it did not come from CompTIA.
How to apply this when you practise
The scoring model changes what "practising" means:
- PMI-style (all-or-nothing per item): a partial answer is worth nothing, so the question is not "can I get two of these" but "which set is the complete set". Practise reading the stem for what it excludes.
- ISC2-style (compensatory): a weak domain is recoverable, so a diagnostic by domain is the highest-value thing you can do. It tells you where to spend the remaining weeks.
- CompTIA-style (scaled, undisclosed): treat the objective list in your report as the definitive map of your gaps, and use the free 100-900 scale as a progress number rather than a pass/fail proxy.
Is there partial credit on PMP multiple-response questions?
PMI does not publish an explicit statement about partial credit. What it does publish is the scoring rule — one point per scored question — which makes each item all or nothing. Present that as what the rule implies rather than as a quotation.
Does ISC2 give partial credit on its advanced item types?
ISC2 does not publish a statement about partial credit on drag-and-drop or hot-spot items. The published rule is a single pass/fail result on the total of all operational items, with compensation across domains. Do not assume item-level partial credit.
Is there a penalty for answering wrong on CompTIA?
No official CompTIA source states either way, so it should not be asserted as fact. For PMI and ISC2 the published models contain no penalty term: PMI awards one point per correct scored question, and ISC2 computes a single result over the operational set.
How close to the pass mark do I need to be?
For PMI you cannot know — the cut is psychometric and the report is a band, not a percentage. For ISC2 you also cannot know: you get no number, only a pass or fail, with domain levels if you failed. For CompTIA the scale is public (750 of 100-900 for Security+), but the mapping from questions to scaled score is not. Practise against domain-weighted mocks and assume you need to be comfortably past the line, not near it.
Related Articles
Explore more articles on related topics

PMP vs CAPM: Which Certification Comes First (and What the Second One Actually Costs You)
6 min read
Online vs Test Center: Which Exam Format Should You Pick? (The Rules Differ by Vendor)
6 min read
ITIL 4 Foundation: The 7 Practices That Are Almost Half the Exam
5 min read
The PMP Exam Changed in July 2026: New Domain Weights, New Question Types
5 min readRecommended Certifications
Based on the topics covered in this article
Ready to Master Your Certification?
Join thousands of professionals who are passing their exams with confidence using our adaptive engine.



