What are training evaluation survey questions?
Training evaluation survey questions are the items you ask learners before, during and after a course to find out how they experienced it, what they learned, whether they used it at work, and what changed as a result. Each question answers one of those four things, and each has a right moment.
This guide is for trainers and learning teams who want more than a satisfaction score. The 24 questions are adaptable examples, not a validated scale, and one fictional training team with 40 completers shows how to read the answers.
THE SHORT VERSION
- Ask each question when its evidence can exist: reaction at the end of a session, learning at exit, use at work after 30 days, results after 90.
- Ask the baseline and exit questions in identical wording, and match them on one learner ID from registration, so you measure change for each person.
- Report every result with its base and its unknowns, and read the open answers for the reasons behind the numbers.
What are good training evaluation questions? 24 examples by when to ask
Good training evaluation questions ask about one thing, at the moment the evidence can exist, in wording you can repeat later. The groups follow the four levels of the widely used Kirkpatrick model (reaction, learning, application or behaviour, and results), plus a baseline before training. Replace the brackets with your course’s skill and choose a few from each group.
Before training: baseline (questions 1 to 4)
Ask at registration or before the first session.
| # | Question | What it tells you | Answer type |
|---|---|---|---|
| 1 | How confident are you doing [the skill] today? | Starting point for the same question at exit | Scale 1 to 5 |
| 2 | How often does your current role ask you to do [the task]? | Whether there will be a chance to use it | Frequency scale |
| 3 | What do you most want to be able to do after this course? | The learner’s own goal, to revisit at exit | Open text |
| 4 | What might make it hard to use new skills at work? | Barriers you can plan for early | Open text |
The five groups sit on one learner’s timeline, shown here for Maria, ID 0417, from this guide’s fictional cohort.
End of session: reaction (questions 5 to 9)
Ask before people leave. Reaction tells you how to deliver the course, not whether it worked.
| # | Question | What it tells you | Answer type |
|---|---|---|---|
| 5 | How relevant was today to your work? | Whether the content matched the job | Scale 1 to 5 |
| 6 | Which part of today was most useful for your work? | What to keep | Open text |
| 7 | Which part needed more explanation or practice? | What to fix before the next session | Open text |
| 8 | How clear were the materials and examples? | Delivery quality | Scale 1 to 5 |
| 9 | What one change would make the next session better? | A specific, usable fix | Open text |
Exit: learning (questions 10 to 14)
Ask at the last session, and pair the self-rating with a task the learner can show.
| # | Question | What it tells you | Answer type |
|---|---|---|---|
| 10 | How confident are you doing [the skill] today? (same wording as question 1) | Change for the same person | Scale 1 to 5, unchanged |
| 11 | Show us: complete [a short task] using the agreed checklist. | Demonstrated skill, not a self-rating | Scored by the trainer |
| 12 | What are the main steps in [the procedure]? | Recall of the core content | Short answer, scored |
| 13 | How would you handle this situation: [a realistic case]? | Whether knowledge transfers to a case | Open text, scored |
| 14 | Which part of the skill still feels unclear? | Where learning is incomplete | Open text |
30 days: application (questions 15 to 20)
Ask once people have had a real chance to use the skill, and ask about opportunity first.
| # | Question | What it tells you | Answer type |
|---|---|---|---|
| 15 | Have you had a chance to use [the skill] at work since the course? | Opportunity, kept apart from use | Yes / not yet |
| 16 | Have you used it at work? | Application | Yes / no |
| 17 | Describe one time you used it. What happened? | Evidence of real use | Open text |
| 18 | What got in the way? | The reasons behind a “no” | Open text |
| 19 | What helped you apply it? | Support worth repeating | Open text |
| 20 | What support would help you use it more often? | What the organization can change | Choice plus open text |
90 days: results (questions 21 to 24)
Ask when results could plausibly show, and look for evidence beyond opinion.
| # | Question | What it tells you | Answer type |
|---|---|---|---|
| 21 | What has changed in your work since the course? | The learner’s view of results | Open text |
| 22 | Which records could confirm that change, such as an error log or case file? | Where to check beyond opinion | Open text |
| 23 | What else changed in your work or team over the same period? | Other influences on the result | Open text |
| 24 | For the manager: what change have you seen in this person’s work since the course? | A second view of the same person | Scale plus open text |
The CDC’s training guidance draws the same line between learning and its transfer to the job. Read the CDC guidance.
What should a short training evaluation form include?
A short form serves one decision, with four to six questions and at least one open question that asks why. Cut anything that will not change what you do, and offer “no opportunity yet” wherever that is a real answer.
Here is a six-item exit form for a one-day course, built from the list above.
SAMPLE EXIT FORM · SIX ITEMS
- How relevant was today to your work? (1 to 5)
- Which part was most useful for your work?
- Which part needed more explanation or practice?
- How confident are you doing [the skill] today? (1 to 5, as at registration)
- Show us: complete [a short task] using the agreed checklist.
- What might make it hard to use this at work?
The 30-day follow-up can be shorter still: questions 15 to 18. Keep post-only items such as relevance apart from the questions you compare before and after.
How do you pair pre and post training survey questions?
Ask the same question, in the same words and on the same scale, before and after training, and compare each person with themselves. Questions 1 and 10 above are a pair; change one word or one scale point and the comparison partly measures your edit.
Pairing also needs to know whose answers are whose. Two piles of anonymous answers give two averages, and the average can rise because different people replied. Match on an ID given at the first form, usually registration, not on a typed email.
Read it one person at a time: 0417 rose two points, 0592 stayed at 3, 1103 rose one, and 0788’s change is unknown, not zero. Report “2 of 3 matched learners rose,” with the fourth shown as missing.
In Sopact Sense, the ID is issued at registration and every later survey lands on the same learner record, so nobody matches files afterwards. Lesson 5 of the free course, Measure change, turns a pair like this into an outcome measure you can defend.
The short video below makes the same case; watch for the point where disconnected metrics are set against one participant ID. For the full method, see pre and post surveys.
Worked example: what did the training team learn at 30 days?
Of 40 fictional completers, 25 answered the 30-day follow-up: 15 had used the skill at work, 10 had not, and 15 outcomes are unknown. The team runs skills training for employers and wanted to know what stopped the others.
Maria’s record shows one full thread: confidence 2 out of 5 at registration and 4 at exit, 10 of 12 sessions attended, a mentor note that she led a mock interview, and at 30 days, the skill used on the job. Each answer sits on ID 0417, dated.
The 10 who had not used it answered question 18. One wrote, “My manager moved me to the evening shift and there was nobody to practise with.” Read together, the answers grouped into three reasons: a changed schedule, no chance to practise, and a manager who had not asked for the skill.
None of those reasons says the course failed to teach, so the team decided to test a practice session with the next cohort instead of rewriting the content.
How do you interpret training evaluation responses?
Report every figure with its base, keep non-respondents visible as unknown, and read the open answers before deciding what a number means. “60% used the skill” is true only of the 25 who replied. The honest sentence is “15 of 25 respondents used the skill at work; 25 of 40 completers replied; 15 are unknown.”
The unknowns set a range: between 15 and 30 of the 40 completers used the skill. If the silent 15 attended fewer sessions than respondents, the 60% flatters the cohort. The deep dive Who stops responding shows how to follow up without pressuring anyone.
Read question 15 before question 16: a learner with no chance to use the skill has not failed at anything. When a manager’s view differs from the learner’s, record both.
Theme each open answer once and keep every theme linked to its answers, so a reader can open the quotes behind “no chance to practise”. In Sopact Sense, an Intelligence Cell applies a prompt your team configures to each open answer as it arrives, and field selection keeps names and emails out of what the AI model receives.
Know the limits: answers are self-reported, the 15 silent learners may differ from those who replied, and Maria’s rise from 2 to 4 does not show the course caused it; that claim needs a comparison group planned in advance. People decide what a theme means, and a barrier comment should never become a negative label on an employee.
How do you start with one course?
Pick one course with a cohort starting soon, choose eight to ten questions from the list, and run them from registration to the 30-day follow-up on one learner ID.
- Write down the one decision the evaluation should support.
- Choose one or two questions from each group, and copy the baseline confidence question word for word into the exit form.
- Give each learner an ID at registration and make every later survey carry it.
- Send the 30-day follow-up to every completer, so non-respondents stay in the count as unknown.
- Theme the open answers as they come in, trace each theme to its quotes, and have a second person check them.
- Meet with the trainer and one manager, agree one change, and test it with the next cohort.
After the first cycle you should have a matched change for each learner, a 30-day result with its base and unknowns, barrier themes linked to real answers, and one change to test.
Frequently asked questions
What is a training evaluation questionnaire?
It is a set of questions given to learners, and sometimes their managers, to judge a course against its objectives. A good one covers more than satisfaction: reaction at the end of a session, learning at exit, use at work after about a month, and results later, each asked when its answer can exist.
What are pre and post training survey questions?
They are matched questions asked before training and again at the end, in the same words and on the same scale, so you can measure change for each learner. A confidence rating and a task scored against the same checklist are common pairs. Match the two answers on a learner ID from registration, not on a name or email.
How many questions should a training evaluation survey have?
Fewer than most teams start with: four to six items at the end of a session, and three or four at 30 days. Choose from the 24 by the decision you need to make, and include one open question that asks why. Pilot the form with two or three learners and cut any question whose answer would not change what you do.
When should I ask about behaviour change after training?
After people have had a realistic chance to use the skill, usually around 30 days, and again at 90 days for results; at the end of the course you only capture intentions. Ask first whether an opportunity came up, then whether they used the skill and what got in the way, so “not yet” is not counted as failure.
Can learners rate how much they learned?
They can report how much they think they learned, which is useful but is not a skills check. Confidence can rise without skill, and some capable learners rate themselves low. Pair the self-rating with one task they can show at exit, scored against an agreed checklist, and keep both answers on the same learner record.
Can a survey prove business results came from training?
Usually not on its own. A survey shows that learners report change and points to records worth checking, but new managers, tools or targets change at the same time. Ask what else changed, look at records beyond opinion, and plan a comparison group in advance if you need a causal claim.
