Crisis management survey: 16 questions for after the event
A questionnaire for the people who were in it, bounded to one named incident or exercise. Sixteen questions, about six minutes, and a number at the end to compare with the next one. Answer it below, or open it in the editor and put your own incident name in.
- 16questions
- about 6 minto answer
- 5-point agreeplus two open answers
- 1 eventper response
What a crisis management survey measures
Four things a response can fail at, asked of the people who were in it, about one named event.
One event, not a mood. Question 1 asks which incident or exercise you are answering about, and everything after it is bounded by that answer. That is what separates this from a readiness poll: how prepared the organisation feels is a guess, whether the first alert arrived is an observation. It also means the same sixteen questions work after a tabletop exercise and after the real thing, which is how FEMA treats the two.
Four domains, kept apart. Knowing what to do, getting the word, being able to act, and where events differed from the plan. They fail separately and different people fix them: a role nobody knew is a briefing problem, an alert that never arrived is a distribution list, a late decision is an authority level. One item asking whether the response went well tells you none of that.
What it is not. Not a wellbeing check, and not the record of the lesson itself, which belongs with a lessons learned survey template once the review has decided what the lesson was. Every question here is already loaded in the online survey creator, so changing the wording is faster than copying it.
The 16 crisis management survey questions
Every question, the scale it takes, and what a low score points at. Copy one block or the whole set.
The event you are reviewing
Two fields first. An answer that does not say which incident it describes cannot be filed, and an onlooker should not be averaged with the person who ran it.
- Which incident or exercise are you answering about?The name the incident log uses. If it was an exercise, say so here, so you can keep those answers apart.
- What was your part in it?The only segment this survey carries. A part in THIS event is not a job title, so it identifies nobody in a small team.
Did you know what to do
The part of a response settled before anything happens. Role clarity, finding the plan and knowing who decides fail independently, and one item asking whether "the plan worked" cannot say which.
- I knew what my own role was as soon as this was declared.Role at the moment of declaration, not in hindsight. Somebody who worked it out later should be disagreeing.
- I could find the plan or the instructions I needed without asking somebody.A plan nobody can open is a plan nobody has. Low scores point at where the document lives and who has the password.
- I knew who was making the decisions at each stage.At each stage, because handovers are where this breaks. HSEEP asks evaluators to record "what key decisions were made", and nobody can produce that record if nobody knew the decider.[1]
Did the word reach you
The domain most crisis surveys reduce to one satisfaction item. These three separate the alert, the updates and the route back, because a response can broadcast well and still leave somebody unable to send anything upward.
- The first alert reached me through a channel I actually check.A channel they check, not a channel you sent to. This is the item that finds the distribution list nobody has updated.
- The updates told me what to do next, not only what had happened.An update that describes events without saying what changes for the reader is a disagreement here, which is what separates this from asking whether updates were frequent.
- I knew where to send information I had that other people needed.The upward route, asked separately: a good cascade with no return path scores well on every other communication item.
Could you actually act
Whether the response was possible for the person in it: the means, the decision they waited on, the route for raising a problem out of turn.
- I had the access and the permission needed to do my part.Access and permission together on purpose: the login without the authority and the authority without the login both stop the same person.
- Decisions I was waiting on arrived soon enough to be useful.Soon enough to be useful, not soon in the abstract. A decision that arrived after its window counts as one that did not arrive.
- I could raise a problem without waiting for the next scheduled update.The escalation valve. Research on incident reporting in health care describes fear of the consequences as a barrier to speaking up, so this asks about possibility rather than procedure.[5]
Where it differed from the plan
The two items that turn a satisfaction survey into an evaluation. HSEEP has its evaluators ask "What happened? What was supposed to happen based on current plans, policies, and procedures?" and "Was there a difference?"[1]
- Something happened during this that our plan did not cover.Agreement here is a good outcome: somebody saw a gap and can name it. Read it beside question 13, not inside the average.
- Which of these got in your way during this incident?Tick all: a person blocked three times should not pick one. "Nothing got in my way" is first so a clean response is one tap.
In your words, then the score
Two open answers, then one rating. A rating asked before somebody has thought about the event is a mood reading; asked after, it is a judgement.
- What worked, and should be kept exactly as it is next time?What to keep, before what to change. A review that only collects failures reads as a complaint form.
- What one change would have made the biggest difference on the day?One change, not a list. That constraint is what makes forty of these answers rankable.
- Overall, how well was this incident handled?The one number this survey produces, asked last so everything above it has been thought through. It asks about the handling, not the organisation.
The twelve agree items share one scale: Strongly agree, Agree, Neither agree nor disagree, Disagree, Strongly disagree. It is written out because a bank printed without its answer options is not a survey anybody can send.
One swap-in. Handled across several sites, add a site field after question 2.
Scoring it, and what the number is worth
One headline number from question 16, three domain averages under it, and no external benchmark.
The headline. Count the answers to question 16 and take the share who said very well or well. Paste your counts in below.
The three domain averages. Score the agree items 5 down to 1 and average each block: questions 3 to 5, 6 to 8, and 9 to 11. Question 12 is averaged with nothing, because agreement on it is a good sign. Three numbers and a headline is the whole scorecard.
There is no benchmark, and be careful with anybody offering one. No comparable distribution is published for this. What the score is good for is the same organisation twice: this incident against the last one, this year's exercise against last year's. The first run is a baseline.
| If this is the low block | What it points at | Who can change it |
|---|---|---|
| Knowing what to do (Q3 to Q5) | Roles that exist on paper and are not known at declaration, or a plan people cannot open | Whoever owns the plan and the briefing |
| Getting the word (Q6 to Q8) | An out-of-date distribution list, updates that never say what changes, or no route back upward | Whoever runs alerting and writes the updates |
| Being able to act (Q9 to Q11) | Access held by the wrong people, or an authority level that sends routine decisions upward | Whoever sets permissions and delegated authority |
| Q12 agreed with, or one option dominating Q13 | A scenario the plan does not cover, or one blocker nearly everybody hit | Whoever maintains the scenario list |
Sending the crisis management questionnaire
Three decisions, then the two pieces of text that go out with it.
Send it once the response has stood down, and close it inside a week. FEMA holds the equivalent conversation, the player hotwash, "immediately following the conduct of an exercise", and uses it to give people "time to fill-out Participant Feedback Forms".[1] A survey does not need the room, so it can wait for the stand-down but not much longer: the order things happened in is the first detail to go, and it is the detail this survey captures.
Ask a wider list than the response team. Everybody involved, affected or on standby. The people who kept normal operations running are the ones who can answer question 6 honestly, and the ones left off the invitation.
Settle anonymity before you write the invitation. Anonymous is the default, and question 2 carries the only segment safe in a small team. If you need to route findings to a site or a shift, add that field and say so in the invitation. What destroys a second round is promising anonymity and then quoting a free-text answer only one person could have written.
Do not run this as a search for who was at fault. Work on learning from failure in health care describes psychological safety as what makes open reporting possible at all,[4] and a review used once to attribute blame collects nothing the next time. For a wider read on whether people feel able to speak up, a confidential employee feedback survey is the instrument.
The confidentiality statement to put at the top
How your answers are handled Answers are not linked to your name or your account. The only thing recorded about who answered is the part you played, which you choose in question 2. A role group returning fewer than five answers is reported inside the whole, not on its own. Free text is quoted only where it names no person, and no answer is used in any decision about an individual.
Turning the answers into corrective actions
The review produces a short list with a name and a date on every line. Anything less is a filed document.
Read the two open questions first. The text is where the specific thing lives. Sort question 15 into repeats and singletons before you look at any average: five people naming the same missing access is a corrective action, and the numbers tell you how widely it was felt.
Write the finding as a difference, not a score. HSEEP describes a strong observation as carrying "a clear and direct statement of the issue identified, a brief description of the analysis, and the impact or result of the issue".[1] "Communication scored 3.1" is none of those. "The night shift got the first alert 40 minutes after the day shift, because the rota list is maintained separately" is all three.
Every action gets an owner and a date, agreed in the room. FEMA's after-action meeting exists to reach "final consensus on strengths, areas for improvement, draft corrective actions, concrete deadlines, and owners/assignees".[1] Two or three actions that close beat fifteen recorded.
Send the list back to everybody who answered. It is the only part respondents can see, and it decides whether the next review gets answers. Where the finding is a rule people did not know rather than a system that failed, an employee compliance survey tells you whether the fix landed.
Then pick the adjacent check. An alarm or an evacuation route points at the fire safety survey questions. An email or a password points at the cyber security awareness survey. And if what it exposed was how the standing group works rather than how it responded, use the team effectiveness survey questions.
Crisis management survey FAQ
What is a crisis management survey?
A questionnaire sent to the people who were in an incident, asking what they knew, what reached them, what they could do and where events differed from the plan. It is bounded to one named event, and that boundary is the point: a readiness poll collects opinions about a hypothetical, this collects observations.
When should a crisis management survey be sent?
While people can still remember the order things happened in. FEMA holds the equivalent conversation, the player hotwash, "immediately following the conduct of an exercise".[1] A survey does not need the room, so send it once the response has stood down and close it inside a week. The reason to hurry is detail: the sequence of who told whom is the first thing to blur.
Should a crisis management survey be anonymous?
Anonymous by default, with one exception. What you need candour about is decisions, delays and who was not reachable, and research on incident reporting in health care describes fear of the consequences as a barrier to speaking up.[5] The exception is routing: if you need to know which site had no access, add a site field and say so in the invitation. Question 2 carries the only segment safe in a small team.
Is this the same thing as an after-action review?
No, it is the input to one. The review is the meeting; this is what you take into it. FEMA describes the equivalent forms as data that "helps to generate the AAR/IP", and lists feedback forms and surveys among the collection methods available to an evaluator.[1] It is also why fourteen of these sixteen questions are closed.
Does running the review actually change anything?
A meta-analysis of 46 samples covering 2,136 people found debriefs "improve effectiveness over a control group by approximately 25% (d = .67)",[2] and a later meta-analysis of 61 studies covering 915 teams and 3,499 individuals put the overall effect at d = 0.79.[3] Both measured the review conversation itself, in training and work settings, not a survey feeding one, so read them as the case for holding the review rather than a result this questionnaire promises.
Can we use this after a tabletop exercise instead of a real incident?
Yes, and question 1 is where you say which it was. FEMA evaluates an exercise and a real-world incident through the same chain, so the items do not change. Keep the two sets apart: an exercise ends when somebody says it does, so its answers on waiting for a decision mean something different.
Michael Hodge, survey methodology and questionnaire design · Updated 9 September 2026 · How templates are reviewed
Sources (7)
- US Department of Homeland Security, Federal Emergency Management Agency. Homeland Security Exercise and Evaluation Program (HSEEP), January 2020. fema.gov
- Tannenbaum SI, Cerasoli CP. "Do team and individual debriefs enhance performance? A meta-analysis." Human Factors 2013;55(2):231-245. 46 samples, N = 2,136. doi.org
- Keiser NL, Arthur W. "A meta-analysis of the effectiveness of the after-action review (or debrief) and factors that influence its effectiveness." Journal of Applied Psychology 2021;106(7):1007-1032. 61 studies, 915 teams and 3,499 individuals. doi.org
- Edmondson AC. "Learning from failure in health care: frequent opportunities, pervasive barriers." Quality and Safety in Health Care 2004;13(Suppl 2):ii3-9. doi.org
- Smit C, Peddle M. "Experiences and Perceptions of Registered Nurses Who Work in Acute Care Regarding Incident Reporting: A Scoping Review." Healthcare 2025;13(11):1250. doi.org
- International Organization for Standardization. ISO 22301, Security and resilience: Business continuity management systems. Named and linked only. iso.org
- American Association for Public Opinion Research. "Best Practices for Survey Research." aapor.org
Sixteen questions written for this page, ordered as the chain a response runs through and following the sequence FEMA evaluators work in: what was supposed to happen, what happened, was there a difference, what was the impact.[1] Twelve share one worded five-point agree scale so the domains can be averaged, and the open answers sit before the closing rating. Wording and ordering follow published practice on clarity and bias.[7] No item here comes from a licensed or standards-body instrument. The FEMA HSEEP doctrine[1] is a US federal work, quoted at a few sentences with attribution; it contains no questionnaire, so there was nothing to reproduce. ISO 22301 is named and linked as background only, and no claim about its contents is made.[6]
Run the review while people still remember the order
Sixteen questions, about six minutes, and one number to compare with the next incident. Put your own incident name in and send it.
Use this template