Post Training Survey Template
A post training survey works twice. Fourteen questions go out while the room is still full. Six different questions go out three to four weeks later, once people know whether they used any of it. Take the same-day survey here, copy either set, or open it in the editor.
- 14questions in the room
- About 4 minto complete
- 2 wavessame day, then 3 weeks on
- 5-pointagree scale
When to send it: in the room, or three weeks later?
Two sends, three weeks apart, with almost no questions in common. The moment decides what a training survey can honestly ask.
Check you are on the right page first. Is the feedback about a named trainer rather than about the session? The facilitator feedback survey is the one built for rating a person. Do you want a before-and-after reading? Take a baseline with the training presurvey first. Does training here mean someone's first ninety days? Then the induction feedback survey is the cadence you want.
Day one belongs to the session. The answers you want are about things people can still see: the pace, the practice, the examples. An hour later all of that has flattened into "it was good".
Week three belongs to the work. On the day, nobody has used anything yet, so a question about use is a prediction. Three weeks on, a question about the pace of the slides is unanswerable. That is why the two waves share almost no items.
The trigger is the delivery date, not the calendar. Set the second wave from the day that cohort ran. A monthly training survey averages three cohorts at three different distances and tells you nothing about any of them.
| Wave | What it can honestly ask | What triggers it | Length |
|---|---|---|---|
| Same day, in the room | What the session was like, and what people think they can now do. | The last slide, before anyone stands up. | 14 questions, about 4 min |
| Three to four weeks on | Whether the thing was used, and what stopped it. | A reminder you set on the day you delivered. | 6 questions, about 2 min |
| One survey a month later | Neither well: the room is a blur and the barriers have not been named. | A gap in the diary rather than a decision. | Avoid |
The 14 post training survey questions, and the 6 that follow three weeks later
Every question, the scale it uses, and what a bad answer to it looks like. Copy one block, or all twenty at once.
In the room, block 1: was it worth it, and can you use it
The four items that carry the most signal. What people found useful travels further than what they enjoyed.
- What we covered applies directly to the work I do.A low answer is a booking problem, not a teaching problem.
- This session was worth the time it took.The item budget holders read. Low means the session gets shorter or gets dropped.
- I can name at least one thing I will do differently because of it.High here with a blank box at question 12 means people are being polite.
- I am confident I can do what this session taught, without help.Read this one first. Confidence predicts both what people learn and whether they use it; being impressed predicts neither.[1]
In the room, block 2: how the session actually ran
The only block you can change before the next cohort. Each item points at one fixable thing.
- The material was explained clearly enough to follow.Low points at the first twenty minutes, not at the audience.
- The session moved at a pace I could keep up with.Read the spread. Too fast for half the room and too slow for the rest averages to fine.
- There was enough time to practise, not just watch.Practice time is the cheapest thing on this list to change before the next cohort runs.
- I can find the materials again when I need them.A low answer today becomes a low use answer in three weeks.
- What would have improved this session most?Whatever most people pick is the change. You do not need a meeting to decide it.
In the room, block 3: what happens on Monday
Predictions, not measurements. They are here to set up the second wave and to show who needs a nudge.
- I know what to do first with this when I am back at work.This is what the follow-up wave gets set against, so a vague answer is worth chasing before people leave.
- I will have a chance to use this within two weeks.Low is not a training failure. It means the course was scheduled too early.
- What is the one thing you will try first, and when?Read these before any average. Blank boxes under high ratings are the clearest warning here.
- What got in the way of learning today, if anything?Room heat, missing logins and a clashing deadline all turn up in this box.
- Overall, I would recommend this session to a colleague doing my job.Last on purpose. Asked first, it colours every answer after it.
Three to four weeks later: the follow-up wave
Six questions sent as a separate survey. Nothing here can be asked on the day, and nothing from the day is repeated. This wave answers best when the last one visibly changed something, so open every session by naming the change the previous cohort asked for.
- Since the session, have you used what it taught?The wave hangs on this. Its three no options separate a training problem from a work problem.
- I can do this now without looking anything up.The honest read on learning, taken once the good mood has worn off.
- What has stopped you using it most?An answer naming a manager or a missing tool is not a training fix.
- Describe one time you used it. What happened?One specific story is worth more than the whole rating column.
- I have had the support I needed to put this into practice.Low across a cohort points at the managers, not at the course.
- Looking back, the session prepared me for the real thing.Set it against the day-one rating. The drop is the most useful number this survey gives you.
| Choice question | The five options |
|---|---|
| In the room: what would have improved this session most? | More time to practise; more time on the hardest part; fewer topics covered properly; better material to take away; nothing, it was right as it was |
| Follow-up wave: have you used what it taught? | Yes, several times; yes, once; no, the work has not come up; no, I tried and it did not work; no, I had forgotten too much |
| Follow-up wave: what has stopped you using it most? | No chance to use it yet; my manager wants it done the old way; the tools or access are not set up; I cannot remember enough; nothing has stopped me |
All twenty questions are already written. The button on this page opens the same-day wave in the editor, with its fourteen questions in place, so you can create a free survey and change any wording before you send it. The follow-up wave is a second survey you make yourself: copy the six-question block above and paste it into a new one. The two waves share no items, so there is nothing to duplicate.
Which level of training evaluation does each question reach?
Four levels, named in the order most people learn them, with an honest note on where this instrument stops.
| Level | What it asks about | Which of these questions reach it | What it cannot tell you |
|---|---|---|---|
| Level 1, reaction | What people made of the session while it was still in front of them. | Questions 1 and 2, questions 5 to 9, and question 14. | Whether anybody learned anything. |
| Level 2, learning | What someone can now do or say that they could not do before. | Questions 3 and 4, as a self-rating. | What they can actually do. A rating is not a test. |
| Level 3, behaviour | What changes back at the desk once the room has emptied. | Questions 10 and 11 predict it. The follow-up wave measures it. | Anything at all on the day. Nobody has been back to work yet. |
| Level 4, results | What moved in the business because the behaviour changed. | None of them. | It is not a survey question. It comes from the numbers the training was meant to move. |
The four levels were set out by Donald Kirkpatrick in 1959 and are now held and developed by Kirkpatrick Partners LLC.[2][3] The descriptions above are ours, written for this survey. The two open boxes, questions 12 and 13, sit outside the levels: they collect the reason behind a rating rather than reaching a level of their own.
Almost every published list of training questions sits entirely at level 1. That is not laziness. It is the only level a survey handed out in the room can reach. Naming the levels is how you stay clear about where the instrument stops.
Level 2 asks what someone can now do, and a self-rating is the weakest way to find out.
What a high score on the day is actually worth
The same-day answers tell you about the session. They do not tell you the training worked, and saying so is the most useful thing this page does.
What the evidence says. A funded research summary of the meta-analytic work is blunt about this. Reaction scores relate only modestly to what people knew straight afterwards. They do not relate at all to what they still knew later.[1] Whether people liked the session does not predict use on the job at all. Whether they thought it would be useful predicts it only weakly.[1]
The number, if you want one. Will Thalheimer puts the average correlation between smile-sheet ratings and learning at about 10 percent across four meta-analyses.[4] The plain version is better: a high score on the day is not evidence.
One item does carry signal. Confidence that you can do the thing predicts both what people learn and whether they use it.[1] That is why question 4 is in the survey, and why it is worth reading before the overall rating.
There is nothing to benchmark against. No widely used measure of trainee reactions exists, so every organisation writes its own.[1] A published band for a good training score is somebody else's house average against somebody else's questions.
A cohort is small, so read the spread. On nine responses, one person moving from 2 to 5 lifts the average by a third of a point. Compare the shape of the answers and the improvement option people picked, not the mean.
What the day-one survey is for. It is process evaluation: finding what to change before you run the session again.[1] Treat it as a maintenance instrument, not as proof that the money was well spent.
How to run both waves without losing the second one
Six decisions that belong to a two-wave training survey and to nothing else. The mechanics every survey shares are in the survey research guide.
Put the link on the last slide, not in a follow-up email
Give people the four minutes before you close the session. A survey that leaves the room with them is competing with the rest of their inbox.
Give every cohort its own link
Duplicate the survey for each session so the results stay apart. There is no session question in the instrument and you do not want one.
Book the follow-up wave on the day you deliver
Put the second send in the calendar before you leave the room. It is the half everyone skips, and three weeks later nobody remembers.
Freeze the 14 items and change only the options
Course-specific wording goes in the improvement options and the open box. Everything else stays identical, or sessions stop comparing.
Name a survey owner who was not in the room
Whoever owns this survey reads the individual answers. Print that name on the last slide, and never let it be the person who ran the session.
Act on the gap between relevance and confidence
High relevance with low confidence means the session was interesting and did not land. The fix is practice time, not more material.
Post training survey FAQ
When should I send a post training feedback survey?
Twice. Hand the first survey out before people leave the room, while the session is still in front of them. Send the six-item wave three to four weeks later, once there has been a chance to use it. One survey a month later gets you a vague memory and a vague answer.
How many questions should a post training survey have?
Fourteen in the room, six later. In the room you are competing with people packing up, so every item has to earn its place: fourteen is what was left after cutting every question whose answer would not change a decision. The follow-up wave is six because it asks about one thing.
Should training evaluations be anonymous?
Be careful what you promise. Whoever owns this survey sees the individual answers. In a group of nine, someone who was in the room can often guess who wrote what. Call it confidential, name the person who reads it, and make sure that person is not the trainer.
Does a good training feedback score mean the training worked?
No. Reaction scores are only modestly related to what people knew straight afterwards, and unrelated to what they still knew later.[1] Liking a session does not predict whether it gets used at work.[1] Use the day-one answers to fix the next session, and the follow-up wave to find out whether anything changed.
Can I use the same questions for every course?
Yes, and you should. Identical wording is the only way to compare one session against another. Put the course-specific detail in the open question and in the improvement options. Change an item and you have started a new series.
What is the difference between a training feedback form and a course evaluation?
A training feedback form is the sheet handed round after a session, and that is what this template is. A course evaluation runs at the end of a course with a syllabus, and the rating it produces is read by somebody other than the trainer, which changes how it has to be written and read. If that is your situation, the course evaluation survey is the one to use.
Methods and sources
What this template is based on
The rules for reading the answers come from a funded research summary of the meta-analytic work on trainee reactions.[1] The two-wave design is ours: that source is about what reaction scores are worth and says nothing about when to ask. Every question is written for this page.
Question licensing
The four levels of training evaluation originate with Donald Kirkpatrick, and the marks are held by Kirkpatrick Partners LLC.[2][3] Nothing from their publications is reproduced or adapted here, and this template is not certified by, endorsed by or connected to them.
Limits and disclosure
A same-day survey measures reaction, not effect, and the follow-up wave is self-reported. There is no benchmark for a training score, because no standard instrument exists.[1] A cohort is usually too small for an average to be stable, so read the spread.
SuperSurvey makes the editor this template opens in. The questions are free to use with it or without it.
References
- Paul, M. "Umbrella summary: trainee reactions." Quality Improvement Center for Workforce Development, 2022. qic-wd.org
- Kirkpatrick, D. L. "Techniques for evaluating training programs." Journal of the American Society for Training Directors 13, no. 11 (1959): 3-9, as cited in qic-wd.org (PDF)
- Kirkpatrick Partners LLC, the company that holds and develops the four levels today. Company site. kirkpatrickpartners.com
- Thalheimer, W. "Learner survey results are not correlated with learning results." Work-Learning Research, 2019. worklearning.com
Send the first wave before the room empties
Fourteen questions, about four minutes, and a six-question follow-up already written for three weeks from now.
Use this template