Each of the 26 tasks in the exam content outline has a characteristic wrong answer that feels responsible and is not: escalating a vision disagreement, smoothing over a conflict, doing the work yourself, reporting green while the project slips. Learn the trap for each task and a large share of the exam stops being ambiguous.
PMP questions are not hard because the right answer is obscure. They are hard because one of the wrong answers is the thing a reasonable, decent, conscientious person would actually do — and it is wrong anyway.
That wrong answer is not random. It repeats. Each task in the exam content outline attracts a particular kind of plausible mistake, and once you can name the mistake for a task, you can see it in the options before you have finished reading them.
What follows is the full set: one named trap for each of the 26 tasks, plus the two that come with the newer material. They are drawn from our own PMP guide, where each task chapter carries its trap alongside the concept and the practice questions.
Almost every trap on this exam works the same way: it takes a genuine virtue and applies it at the wrong moment.
Escalating is responsible — except when the decision is yours. Keeping the peace is mature — except when the disagreement is the information. Helping is generous — except when it teaches your team to wait for you. Reporting more is diligent — except when nobody uses the report to decide anything.
So the trap is rarely laziness or incompetence. It is a good instinct pointed one step too early, one step too late, or one level too high. That is why candidates who are genuinely good at their jobs still lose marks: their working habits are being tested against a specific model of what a project manager owns.
The remedy is not to distrust your instincts. It is to know, for each task, which good instinct that task punishes.
The People domain carries 33 percent of the exam. Its traps almost all involve doing something to a person that avoids the harder work of doing something with them.
| Task | The trap | Why it is wrong |
|---|---|---|
| Develop a common vision | The escalation that looks like leadership | Building and holding the shared vision is your own job; sending it upward buries a solvable misalignment |
| Manage conflicts | The peacekeeper | Smoothing it over leaves the real disagreement alive to resurface later, at a higher cost |
| Lead the project team | The rescuer | Stepping in to do the work personally keeps the team dependent and teaches people to wait for you |
| Engage stakeholders | One-size-fits-all engagement | The same report to everyone more often is not engagement — different audiences need different things |
| Align stakeholder expectations | False consensus | No objection is not agreement; nobody has aligned, they have simply not spoken |
| Manage stakeholder expectations | Set-and-forget | Aligned once is not aligned now — this is how projects lose customers without notice |
| Help ensure knowledge transfer | The single point of knowledge | Leaving what one person knows inside that one person |
| Plan and manage communication | Status theater | More emails, more charts, bigger packs — activity mistaken for information |
Process is the heaviest domain at 41 percent, and its traps cluster around a single failure: accepting a comfortable number or a comfortable plan instead of the real one.
| Task | The trap | Why it is wrong |
|---|---|---|
| Develop an integrated plan and plan delivery | The default approach | Running the method you always run instead of the one this project needs |
| Develop and manage scope | Scope creep | Each small yes is reasonable; together they blow the scope with no change request and no fresh agreement |
| Help ensure value-based delivery | The feature factory | Shipping output and calling it outcome |
| Plan and manage resources | Assumed availability | Planning as if named people and shared resources were fully and exclusively yours |
| Plan and manage procurement | The lowest bid | Price chosen ahead of capability, risk and total cost |
| Plan and manage finance | The green number | A financial figure reported as healthy without the analysis behind it |
| Plan and optimize quality | Gold-plating | Polish nobody asked for and nobody is paying for, which can introduce fresh risk |
| Plan and manage schedule | The wish date | A date that reflects what someone wants rather than what the work supports |
| Evaluate project status | The watermelon status | Green on the outside, red on the inside |
| Manage project closure | The fizzle-out | Letting a project drift to a stop instead of closing it |
Business Environment carries 26 percent, and its traps share a shape too: the artefact exists, so the work is assumed to have happened.
| Task | The trap | Why it is wrong |
|---|---|---|
| Define and establish project governance | Paper governance | The board meets and the documents are filed, but no decision gets made and no threshold triggers anything |
| Plan and manage project compliance | The tick-box | Boxes checked for an auditor rather than requirements genuinely met |
| Manage and control changes | Just do it | Building the fix directly — no request, no impact assessment, no approval, no updated document |
| Remove impediments and manage issues | Still a risk | The pump failed and flooded a section, and it is still listed as a risk to monitor |
| Plan and manage risk | Register and forget | The workshop happened, the spreadsheet was saved, and it has not been opened since |
| Continuous improvement | The lessons graveyard | Capturing lessons faithfully and then burying them |
| Support organizational change | Culture blindness | A sound plan driven into a wall of how people actually work, which nobody assessed |
| Evaluate external business environment changes | Looking inward | Heads-down and on schedule, while the threat that kills the project arrives from outside |
The July 2026 outline brought AI in project work and sustainability considerations into scope, and each arrived with a trap of its own. Neither has a domain — both ride inside tasks you already know.
The AI oracle: treating an AI's output as a verdict to obey rather than an input to weigh. The tool is fast and sounds certain, so the tempting answer follows it without checking the data, questioning the recommendation, or owning the decision. The accountability never transferred — if the AI is wrong, it is still the project manager's call that failed. Use AI as a very capable advisor whose work you verify.
Greenwashing: claiming or reporting sustainability the project has not actually delivered. The label goes on the marketing, the report shows the target as met, the box is ticked — while the practice underneath does not match the claim. A hollow claim is worse than none: it collapses on inspection, damages trust, and often breaks a compliance rule.
Both follow the pattern of every other trap on this list. Something that looks like the responsible action, standing in for the responsible action.
Naming traps tells you what to avoid. These five decide what to choose, and between them they resolve most questions where two options look defensible.
Understand before you act. Options that gather information or diagnose the cause usually beat options that apply an immediate remedy — unless the scenario has already established the cause, in which case gathering more is stalling.
Let value decide. When two paths are both possible, the one that protects or delivers the value the project exists for is the answer.
Collaborate before you escalate. Escalation is for decisions genuinely outside your authority. Most of what feels escalation-worthy is a conversation you have not had yet.
Own it. Accountability does not transfer — not to a sponsor, not to a committee, not to a tool.
Match the approach to the context. Predictive, agile or hybrid is set by the scenario, not by preference.
One exception applies to the first rule and it matters: where the scenario involves safety, a legal breach or a compliance failure, the exam does not want you to investigate first.
The failure mode here is mechanical. A candidate who has memorised twenty-eight traps starts eliminating any option that resembles one, and eliminates the correct answer along the way — because escalation is sometimes right, and so is stepping in personally, and so is reporting more.
Use them as a first read, not a verdict. When an option matches a trap for that task, ask whether the scenario has established the conditions that make it correct anyway. Usually it has not, which is why the pattern holds. Sometimes it has, and that is the question testing whether you were reading or pattern-matching.
The productive drill is to work practice questions and, for every one you get wrong, write down which trap you fell for. Most candidates find they have two or three favourites rather than twenty-eight — a bias toward escalating, or toward helping, or toward reporting. A bias you can name is a bias you can correct in a fortnight.
Our own guide carries each of these traps in the chapter for its task, next to the concept and the practice questions, and collects them again in an appendix — because the trap is easiest to learn beside the material it applies to, and easiest to revise as a list.
Twenty-six across the three domains: eight in People, ten in Process and eight in Business Environment, weighted 33, 41 and 26 percent respectively on the July 2026 outline.
Premature escalation — sending a decision upward that belongs to the project manager. It appears as a plausible option across the People and Business Environment domains, and it feels like the responsible choice, which is exactly why it works as a distractor.
Yes — when the decision is genuinely outside your authority, typically anything that moves the scope, schedule or cost baseline, or that belongs to governance. The trap is escalating what you could and should resolve yourself.
Reporting a project as green on the outside while it is red underneath. It appears in the Evaluate project status task, and it is tested because the temptation to smooth a status report is universal and the consequence is that nobody can act in time.
No. Neither has a domain of its own. The July 2026 outline weaves both into existing tasks, and each brings its own trap — the AI oracle, and greenwashing.
No, and trying is its own trap. Traps tell you what a wrong answer tends to look like; they do not tell you what the right one is. Candidates who eliminate mechanically eliminate correct answers too, because escalating, helping and reporting are all sometimes right.
Rules change. Where a figure or a procedure can move, the issuing agency’s current published instructions win over anything here.
