A large share of PMP questions never ask you to calculate anything. They describe a situation, then hand you four answer options that differ by a single noun: update the risk register, update the issue log, update the assumption log, update the stakeholder register. The scenario is often easy. Choosing between four artifacts that all sound plausible is where the point is won or lost.
This is a skill you can practice separately from content review, and it pays off across all three domains.
Why this matters more on the updated exam
PMI launched the updated PMP exam on 9 July 2026. Alongside the content changes, PMI describes the updated exam as adding case and scenario questions and graphic-based questions, with more content built around dashboards, artifacts, tools, and data.
The exam format PMI currently publishes is 180 questions in 240 minutes, with two 10-minute breaks. Domain weighting is People 33%, Process 41%, and Business Environment 26%.
Two things follow from that. First, questions built around an artifact are not a niche category you can skip. Second, in a case-study block you may read one scenario and answer several questions from it, so misreading which artifact the scenario is actually about can cost you more than one point.
Three families, three different jobs
Most exam artifacts fall into one of three families. Getting the family right usually eliminates two of the four options immediately.
Logs record things that have already happened or are happening now. They are working documents, updated continuously, and they track items that need resolution or that were recorded as facts.
Registers record identified items and the analysis around them. A register usually holds entries that were deliberately identified, assessed, assigned an owner, and are being monitored. Registers lean forward in time.
Plans and baselines say how the project will be run, or what the approved target is. A plan is changed through the change control process, not by a project manager quietly editing it.
| Family | Typical artifacts | Updated when |
|---|---|---|
| Logs | Issue log, change log, assumption log, lessons learned register (as a running log) | Something has occurred, is being tracked, or has been recorded |
| Registers | Risk register, stakeholder register | Something has been identified and analyzed |
| Plans and baselines | Management plans, scope baseline, schedule baseline, cost baseline | Through approved change requests |
The lessons learned register sits awkwardly between families. Treat it as a running record during the project that is rolled into the organizational repository at closure.
Test one: has it happened yet?
This single question resolves the most common artifact pair on the exam.
If the scenario describes something uncertain — may happen, could affect, there is a possibility — the answer is almost always the risk register. If the scenario describes something that has already occurred — the vendor missed the date, a defect was found, the team cannot proceed — the answer is almost always the issue log.
The trap is the scenario that starts as a risk and then occurs. A supplier delay that was logged as a risk six weeks ago is not a risk anymore once the shipment is late. It becomes an issue, and the risk register entry is closed out rather than updated with the new impact.
Test two: is this a fact, an uncertainty, or a decision?
Three artifacts get confused because the underlying events look similar.
- A statement being treated as true without proof — the vendor will have staff available in Q2 — belongs in the assumption log.
- An uncertain future event with a probability and an impact belongs in the risk register.
- A request to alter approved scope, schedule, or cost goes through a change request, and the outcome is recorded in the change log.
If the scenario says a team member believes or expects something, and the project plan depends on it, you are usually being tested on assumptions, not risks. An assumption that turns out to be wrong then becomes an issue, and may trigger a change request.
Test three: who is allowed to change it?
This test catches the options that are wrong for governance reasons rather than content reasons.
A project manager can update a log or a register directly. A project manager cannot update a baseline directly. If an option says the project manager revises the schedule baseline, updates the budget, or changes approved scope without mentioning a change request, it is almost always wrong, even when it is the practical thing to do.
The same logic applies to anything owned by someone else. The project charter is issued by the sponsor. The project manager does not rewrite it to accommodate a new requirement.
Worked examples
A stakeholder you did not expect appears at a steering meeting and starts raising requirements.
The event has happened, but nothing is yet broken, and the new party needs analysis — influence, interest, engagement approach. This is the stakeholder register, followed by a review of the stakeholder engagement approach. It is not the issue log, because no problem needs resolving yet.
A team member tells you the integration environment will be free from March, so you built the schedule around it.
Nothing has gone wrong and nothing is uncertain in the probabilistic sense — a statement is being relied on. Record it in the assumption log and validate it. If it later proves false, it becomes an issue.
During execution, the customer asks for an extra report that was not in the agreed scope.
A request to change approved scope requires a change request, regardless of how small it looks. The decision is recorded in the change log. Doing the work first and documenting later is the distractor.
A risk you identified in planning has occurred and is now delaying a deliverable.
Close the risk register entry, open an issue, and assess whether a change request is needed. Many candidates pick the risk register because the word "risk" appears in the stem. Read the tense.
Traps that cost the most points
- Matching a keyword in the stem to the same keyword in an option. The exam uses this deliberately.
- Choosing a document update when the scenario calls for a conversation first. If the situation involves a person, the first step is often to engage them, then record.
- Choosing escalation early. Escalation is usually correct only after the project manager has analyzed the situation and found it outside their authority.
- Updating a baseline without approval.
How to practice this
Work in short, focused sets rather than mixed exams. Take twenty questions whose answer options name artifacts, and before you look at the options, write down which artifact you expect and why. Then compare. What you are training is the decision sequence — family, timing, ownership — not recall of definitions.
When you review a question you got wrong, do not stop at the correct answer. Say out loud why each of the other three artifacts was wrong. If you cannot, that distinction is the gap, and it will reappear.
Over a few sessions this becomes mechanical, and a category of question that used to feel like a coin flip turns into reliable points.