feedback forms · Kirkpatrick-ready · corporate training
Training Evaluation Form.folder
A satisfaction survey asks if people liked the day. This evaluation pack asks what changed, what stayed, and what the next run should fix — Level 1 through Level 3, with a report your manager can forward as-is.
// kit v2: a stage-ready HTML deck (presenter drawer, run-of-show timer, built-in QR codes) · poker-spec print-and-cut cards with backs and blanks · an A3 activity poster · manuals in A4 and US Letter with bookmarks · three license tiers (CC BY free / single event / organization).
A full evaluation pack for corporate training: eight reaction items grouped content and delivery, three narrow confidence self-reports as the Level 2 proxy, and two keep/change open questions that become the agenda of the next design meeting. Thirty days later, a three-question email turns feedback into Level 3 behavior evidence — the level most evaluation never reaches.
The props are poker-spec print cards (63.5 × 88.9 mm, bleed and cut guides included) — instruction card, reaction-item cards, confidence card, keep/change card — plus the day-30 email template and the management one-pager in the debrief sheet. Nothing needs an account: hand the cards out in the last minutes, send the email from the address participants already know.
the run sheet
// this is the guide file, word for word — read it here, run from it here. the .md at the bottom of the page is the same text, offline.
Folder:
training-evaluation-form· Category: feedback forms · any group size · 7 min on the day + 3 min at day 30 · paper or onlineLicense: CC BY 4.0
A Kirkpatrick-ready evaluation pack for corporate training: eight reaction items, three confidence self-reports, two keep/change questions — plus the day-30 follow-up that turns feedback into behavior evidence. It produces a one-page report a manager can forward without translation.
What this form covers
Kirkpatrick's four levels, mapped honestly:
| Level | What it asks | Where it lives in this folder |
|---|---|---|
| 1 — Reaction | Did the room find it relevant, deep, well-run? | The 8 items on the form |
| 2 — Learning | Can people now do the core things? | 3 confidence self-reports (a proxy — see below) |
| 3 — Behavior | Is it showing up in the work? | The day-30 follow-up (3 questions by email) |
| 4 — Results | Did business numbers move? | Not covered. Not with a form. Be suspicious of any form that claims it. |
The honest boundary at Level 4 is the feature: evaluation forms that pretend to measure business results produce numbers nobody believes, and then nobody believes the rest of the form either.
Why eight reaction items
Grouped, not piled: four content items (relevance, depth, instructor, materials), three delivery items (exercises, logistics, workload), one overall. In corporate training the delivery items matter as much as the content — a well-taught course with a broken sign-up flow gets scored like a bad course, and the item means will tell you which half to fix.
Level 2: the self-report proxy
The three confidence items are a proxy for learning — cheaper than a test, rougher than a test, and honest about being a proxy. Self-report inflates when the training felt inspiring; it inflates less when the items ask for narrow, checkable abilities ("explain it in two minutes", "apply it without help"). That is why these three are phrased narrow.
If you can collect a baseline — the same three items at sign-up — the delta becomes the credible number. If you cannot, report the absolute values and write "self-report, no baseline" in the report. That one line of honesty is worth more than a chart.
The keep/change questions
The two open questions are not comments; they are the agenda of the next design meeting:
- "What worked so well it should stay exactly as it is?" — protects the good parts from a redesign that "improves" them away.
- "What should change in the next run?" — the fix list, in the participants' own words.
For repeat programs this pair is worth more than every rating combined: ratings tell you where you stand, keep/change tells you what to do.
The day-30 follow-up
Thirty days after the training, send three questions by email (template in the debrief sheet): used-it frequency, what helped or blocked, what support would make it stick. This is the Level 3 behavior check — the level most evaluation never reaches, and the level that answers the question the sponsor actually asked: "did anything change?".
Expect a third of the response rate of the in-room form. That is normal and still enough: behavior counts from a 30% sample are evidence; satisfaction means from a 90% sample are not evidence of behavior at all.
Known failure modes & rescues
- The form becomes a satisfaction contest between vendors. The moment item means are used to rank instructors, answers turn diplomatic and the instrument dies. Rescue: report room-level means to the program owner, never per-instructor league tables.
- The day-30 email never gets sent. The pack halves itself: Level 3 evidence requires the follow-up, and it is always the first thing a busy calendar drops. Rescue: schedule the follow-up email the same day you run the training — write it before you forget who was in the room.
- Self-reports reported as learning gains. Without a baseline, a 4.3 confidence mean is a feeling, not a gain. Rescue: say "self-report, no baseline" in the report — one honest line beats a decorated chart.
Variations
- Multi-session programs: run the form after each session; run the day-30 follow-up once, 30 days after the last session.
- Compliance documentation (Taiwan / Hong Kong): the item means plus the follow-up behavior counts are exactly what quality systems want as training-effectiveness evidence (TTQS course documentation, QF-style records). File the room-level summary; individual sheets stay private.
- Vendor-delivered training: add one item — "My manager knew what I was learning" — the cheapest measure of the transfer climate you will ever install.
the debrief sheet
// the debrief questions, in full — the .md version prints on one page for the room.
One rule first: reaction items describe the day; only the day-30 answers describe the work. Report them separately, never averaged together.
Use with
training-evaluation-form-guide.md. Reading takes 20 minutes for a cohort of 20.
The 60-second read
- Overall item at 4.0 or higher, but relevance under 3.5? The training was enjoyable and aimed at the wrong work. This is the most common corporate failure and the item means catch it precisely.
- Delivery items vs content items. Logistics and workload under 3.0 while content rides high: fix the program administration, not the course.
- Day-30 behavior count above zero? The training is alive. Zero uses in 30 days — the course was an event, not a change, and no reaction score repairs that.
Item-by-item interpretation
| Item | Construct | Low (≤ 2.5) means | First action |
|---|---|---|---|
| 1 | Relevance | Content missed the actual work | Rewrite the needs intake before the next run |
| 2 | Depth | Correct generalities, no depth | Replace one lecture with one worked case |
| 3 | Instructor | Questions met with slides | Instructor prep session on Q&A handling |
| 4 | Materials | Handouts filed and forgotten | Cut to one page people will reopen |
| 5 | Exercises | Listening outweighed doing | Add one exercise per half-day |
| 6 | Logistics | Sign-up, schedule or room friction | Fix the admin flow — it is dragging the scores |
| 7 | Workload | Course load collided with the job | Reschedule or split the cohort |
| 8 | Overall | The room's honest verdict | Read items 1-7 before reacting to this one |
Level 2: reading the confidence self-reports
The three items (apply without help / explain in two minutes / know where to look when stuck) are deliberate narrow abilities. Read them as a set:
- All three low — the course taught awareness, not capability. Not a failure; but call it awareness training in the report.
- "Apply without help" low, the other two high — people understood and remember, but the doing needs support: schedule a practice session, not a repeat of the course.
- With a baseline: the delta is your Level 2 number. Without: report absolutes plus the words "self-report, no baseline".
The day-30 email (template)
Send 30 days after the last session, from the person the participants know, subject: "30 seconds — did any of it stick?":
You attended [program name] on [date]. Three questions, 30 seconds, anonymous:
1. Since the training, I have used the main method: (never / once / 2-5 times / weekly) — circle one
2. What helped or blocked the using? — one sentence
3. What support would make it stick? — one sentence
>
Nothing else. No follow-up will chase you.
The last line is not politeness; it protects the response rate of the next program.
The management one-pager (template)
[Program name] — evaluation summary, [date], N=[n], response [r]%
- Reaction: overall [x.x]/5; lowest item [name] [x.x]; highest item [name] [x.x]
- Confidence (self-report[ + delta]): apply [x.x][ (Δ+x.x])], explain [x.x][ (Δ+x.x])], recover [x.x][ (Δ+x.x])]
- Day-30: [n] of [n] replied; used-it count [n] (never [n] / once [n] / 2-5× [n] / weekly [n])
- Keep: [one verbatim, permissioned] — Change: [one verbatim, permissioned]
- Recommended next run: [one sentence, from the keep/change list]
One page. If it does not fit on one page, the program does not fit in a manager's attention — cut until it does.
3 known failure modes of reading the results
- Levels averaged together. A 4.5 reaction mean next to a zero behavior count is a failed program with a nice cover photo. Rescue: report levels in separate lines — the gap is the finding.
- The one-pager grows a second page. Charts, percentile bands, benchmark logos. Rescue: the template above is the ceiling; anything past it goes in an appendix nobody reads.
- The day-30 obstacles go unread. The follow-up's second question (what blocked the using) is free transfer-climate consulting. Rescue: read those sentences first — they are the program's actual to-do list.
who it's for · what it leaves behind
Who is the Training Evaluation Form for?
L&D owners, HR partners and program managers who need to report training effectiveness to management with something stronger than a satisfaction average. It fits internal programs, vendor-delivered trainings and repeat courses. The item means plus day-30 behavior counts are deliberately shaped to double as training-effectiveness evidence for quality systems (TTQS course documentation in Taiwan, QF-style records in Hong Kong).
What's the difference from the Post-Training Survey?
Scope and audience. The Post-Training Survey is a trainer's one-page instrument — 6 items, 2 open questions, built for speed at the end of any session. This pack is a program owner's instrument: 8 reaction items split content and delivery, a Level 2 confidence proxy, keep/change questions for repeat programs, a day-30 follow-up for Level 3 evidence, and a management one-pager template. Five minutes for one session vs seven minutes plus a month-later email — pick by who reads the report.
Does it really cover Kirkpatrick Level 4?
No — and that is deliberate. Level 4 asks whether business results moved, which requires measurement design outside any form: control groups, operational metrics, time series. A form that claims Level 4 produces numbers nobody believes, and then the honest levels stop being believed too. This pack goes one step past the smile sheet (Level 1) into learning (Level 2 proxy) and behavior (Level 3) and says where it stops.
Can my AI agent use this folder?
Yes. Every folder ships a SKILL.md skill file: frontmatter for matching, plus runnable prep commands — fetch the evaluation cards, pull the reading guide, build the live online version (including the day-30 follow-up as a scheduled second collection) with formlm-cli. The three-step workflow lives on the skills page, and every folder is listed in catalog.json.
make it live
The paper pack is the complete default — cards in the room, one email at day 30, nothing else to buy. If you want the online version — the same 8 items, confidence self-reports and keep/change questions as a live form, with the day-30 follow-up scheduled as a second collection — two ways:
for AI agents
$ formlm-cli smart generate --input "Training evaluation form for corporate training programs: 8 anonymous Likert reaction items (1-5) across relevance, depth, instructor, materials, exercises, logistics, workload and overall value, 3 confidence self-report items (level-2 proxy), 2 open keep/change questions for the next run, plus a 3-item 30-day follow-up email (level-3 behavior check). Produces a management one-pager with item means, confidence deltas and day-30 behavior counts"
Step 4 of the agent workflow: npm install -g @formlm/cli, log in once, and the prep above runs itself. The form and its data live in your own workspace.
— or, with a mouse —
Open the free training evaluation form template on FormLM
and paste the items from training-evaluation-form-assessment.json — the template already collects responses and computes the item means.
// optional, not required — the paper version in this folder is complete on its own. free account on their side.
put it on the wall
The folder ships a single-file HTML deck that presents full-screen from one laptop. To run the activity on the room's big screen — submissions land on the wall in real time, the countdown and the spotlight stay in your hands, and the whole room joins with one scan — the scan-and-project companion is QRCard-AI:
See the trainer live-screen mode on QRCard-AI — one QR code on the wall, participants join and follow from their own seats.
// optional, not required — the deck in this folder already runs full-screen on a single laptop.
take this folder with you
// every file downloads as a plain file. no account, no email wall inside a folder.