What should an employee training survey ask?
Ask questions that help you understand whether training was relevant, what employees learned, whether they had a chance to use it and what happened afterward. Use a few questions at the appropriate stage. An end-of-course survey cannot establish everything you need to know about later performance.
Start with the decision. If you need to improve the next session, ask about the activities and explanations. If you need to support employees applying a skill, ask about opportunities, barriers and recent examples. If you need evidence of learning, include a suitable assessment; confidence alone does not show that someone can perform the task.
The examples below are starting points for an L&D or program team, not a validated questionnaire. Adapt the named task, role and time period, pilot the wording and keep only questions your team will use.
Choose the evidence before choosing the questions
The Kirkpatrick model distinguishes reaction, learning, behavior and results. These help organize an evaluation, but they are not four forms that every employee must complete in sequence. Decide what result matters and what evidence would help you assess progress toward it.
Scroll horizontally to see all columns →
| What you need to understand | Useful evidence | What it does not establish alone |
|---|---|---|
| Reaction | Relevance, experience and immediate concerns. | That the employee learned or will apply the skill. |
| Learning | A knowledge check, scenario or demonstrated task assessed against clear criteria. | That the workplace will give the employee an opportunity to use it. |
| Application | A recent example, observation or work record, with context about opportunity and support. | That training alone caused a later business result. |
| Results | The relevant operational measure, period and other changes that might explain it. | Causal attribution without an appropriate evaluation design. |
A survey may collect part of this evidence. A skills exercise, service record or supervisor observation may supply another part. Keep each source identifiable rather than treating everything as a survey score.
16 employee training questions to adapt
Choose from the following bank rather than sending all 16 at once. Replace “the task” with a concrete activity, such as documenting a service request or preparing a forecast. Label answer choices so respondents know what each point means, and offer “not applicable” or “no opportunity yet” where appropriate.
At the end: relevance and immediate support
- How relevant was this training to the work you currently do? Use a labeled relevance scale, with an option for people whose role does not include the task.
- Which part of the training will be most useful in your work? Use an optional open response. Do not assume every participant found something useful.
- Which part needs a clearer explanation or more practice? Ask about a specific activity if that is the decision you need to make.
- What support would help you try the task at work? Separate a need for practice from missing equipment, permission or time.
These questions can improve delivery and surface immediate needs. They should not be reported as proof of learning. Avoid “How much did our excellent training improve your performance?” because it assumes both quality and improvement.
Before and after: learning and confidence
- How would you handle this scenario? Use a realistic situation and define the criteria for a good response before scoring.
- Demonstrate the task using the supplied case. This is an assessment activity, not a satisfaction question. Record which steps were performed successfully.
- How confident are you that you can complete the task without help? Use a consistent labeled scale. Report this as confidence, alongside performance evidence where available.
- What would you check before making this decision? Evaluate the answer against the intended knowledge or reasoning, not the length or polish of the response.
If you want to estimate individual change, plan comparable baseline and follow-up assessments. Repeating an identical test can introduce familiarity effects; equivalent tasks may be appropriate when designed and checked carefully. A post-only assessment can establish performance against a criterion, but it cannot by itself show how much that person improved.
After a chance to apply the skill
- During the past four weeks, how often did you have an opportunity to use the task? Adapt the period to the work. Distinguish a lack of opportunity from choosing not to apply it.
- Describe one recent occasion when you used it. Ask for enough detail to understand the action without requesting unnecessary personal or customer information.
- What helped you use it? Look for support, access, reminders or practice without assuming one explanation.
- What made using it difficult, if anything? Leave room for a problem with the training and for conditions outside it.
Do not code “no opportunity yet” as failure. A task performed once a quarter needs a different follow-up schedule from one used every day. For employees who have not had an opportunity, decide whether a later check-in or a practice exercise would be useful.
When a relevant result can be observed
- What result did you observe after using the task? Allow “no change,” “too early to tell” and an explanation rather than requiring an improvement.
- What record or observation supports that result? Identify the source and period; collect only evidence the person is permitted to share.
- What else changed during the same period? A new process, staffing change or different workload may matter to interpretation.
- What should the team do next to support effective use? Make the answer useful to the person responsible for follow-up.
An employee's account can explain context and point to evidence. It should not be substituted for an operational measure the employee cannot reliably observe. Check the underlying record before reporting a numerical business result.
When should you send each survey?
Plan the schedule around the skill and the decision. Collect a baseline before exposure where change is the question. Ask about the session while people can recall it. Check application after employees have had a realistic opportunity to use the skill.
A 30-, 60- or 90-day check-in can be useful, but none is a universal requirement. A weekly service task might warrant an early check. A yearly budgeting task may need a later observation. Record the intended window, actual response date and opportunity to apply the skill so that late or early responses remain interpretable.
Let participants know why you are following up, how much effort the survey takes and who will see the results. Coordinate reminders with other employee surveys rather than repeatedly asking for the same context.
Keep the learner and the context connected
For identified evaluation, keep baseline, attendance, assessments and follow-up attached to the appropriate learner ID. Retain the cohort, course version, question version, source and collection date. A changed email address should not create an apparently new learner.
Keep stable registration information in the participant record and update it when needed. Record role or location changes with their timing; today's team should not silently replace the team associated with an earlier response.
Some organizations need anonymous employee feedback. Explain which collections are anonymous and which support identified follow-up. Do not promise both anonymity and person-level matching for the same response, or try to reconstruct identities from comments.
This is the practical reason to plan the record before sending the form. It allows a reviewer to read a reported barrier beside the relevant assessment and later follow-up, within the agreed access rules, instead of rebuilding the story from separate exports.
Can locations ask different training questions?
Yes. A distributed program can keep local questions while agreeing on a small shared core for comparison. Different roles may need different scenarios or examples. Imposing one long questionnaire can add burden without improving the evidence.
Define each shared measure in a data dictionary: what it means, who should answer, response options, scoring, time window and treatment of missing or not-applicable answers. Review whether locally adapted assessments support the comparison you intend to make.
When wording or scoring changes, record the reason and version. Check whether the change affects meaning. If it does, start a new series or clearly limit the comparison; if comparability has been established, document the basis. Neither “never change a question” nor “all similar-looking questions can be combined” is a sound rule.
Analyze change without losing the missing responses
Suppose 100 employees enroll, 80 complete training and 60 completers answer a later survey. Of those 60, 42 report applying the skill. This is an illustrative example.
- Application among follow-up respondents: 42 of 60, or 70%.
- Follow-up coverage among completers: 60 of 80, or 75%.
- Recorded application among all enrollees: 42 of 100. This does not mean the other 58 failed to apply it.
If only 48 people also have a comparable baseline assessment, any paired assessment comparison must use those 48 records and state that coverage. Do not compare the average of one group before training with a different group's average afterward and describe it as each learner's improvement.
Review comments for possible explanations, including accounts that challenge the main pattern. Distinguish respondents who mention a theme from the number of times a theme appears. One person may mention several barriers. Retain source passages so the reviewer can check how a conclusion was reached.
What if your existing questionnaire has 100 questions?
Inventory the questions before adding another form. For each item, record the decision it supports, the measure, who answers, when it is asked and who uses the result. Keep required or methodologically important items; remove duplication and questions nobody can explain how they use.
Check double-barreled wording, unclear time periods, inconsistent scales and assumptions of success. Pilot the revised questionnaire with suitable respondents. Look at misunderstanding and completion effort, not just whether the form technically works.
AI can help organize comments and draft a summary, but a reviewer should check uncertain interpretations, source quotations and calculations. A theme is not an automatic explanation of why a learner's score changed. Access to employee evidence should follow the same rules during analysis as during collection.
Turn the findings into support
Assign a next step to the finding you can substantiate. Unclear instructions may need a revised activity. Lack of equipment may need an operational response. Missing follow-up may need a different collection plan. These are different problems and should not all become requests for another training session.
In a Sopact workflow, plan collection, analysis and governance together: which evidence belongs on the continuing record, what shared definitions guide review and who can inspect or approve the result. Test the intended workflow with a baseline, a later response, a changed question and a missing follow-up before scaling it.
Use the Training & programs course to build that plan. For ready-to-adapt reporting examples, continue to survey report examples. Keep the participant's experience, the evidence and the next decision connected.
Frequently asked questions
How many questions should an employee training survey have?
Use the fewest questions that answer the decision and meet the evaluation requirements. There is no universal number. Pilot the effort and do not combine every stage into one end-of-course questionnaire.
Does a confidence score measure learning?
It measures self-reported confidence. Use a suitable knowledge or skills assessment if you need evidence of performance, and report the two separately.
Must questions stay identical forever?
No. Preserve comparability where it matters, document changes and check their effect on meaning and scoring. Some changes require a new series; others may support comparison with a documented basis.
Is a 90-day follow-up always appropriate?
No. Follow up after a realistic opportunity to use the skill, at a time that supports the decision. Record the window and whether respondents had an opportunity to apply it.
Can a training survey prove business impact?
A survey can contribute evidence, but a self-report alone does not establish that training caused a business outcome. Check relevant records, other changes and the evaluation design.
