Filled example: evaluation chain
Fictional course “Handovers that hold” for 12 shift team leaders on Thursday 12 November 2026. Filled backwards, from the business result to the reaction on the day.
| Level | What we want to see | Measure or sign | Baseline | Target and date | Source and owner | If it is not met, we will… |
|---|---|---|---|---|---|---|
| Level 4: result | Open issues reach the next shift | Logs in a sample of 40 that miss an open issue shown on the incident list | 14 of 40 (35%) in the September 2026 sample | 4 of 40 or fewer (10%) in the February 2027 sample | Quality lead, monthly, against the incident list | Read the logs that still miss an issue and ask whether the cause is skill, time or the log form |
| Level 3: behaviour B1 | Each open issue is logged in four lines: what happened, state now, next action and owner, deadline | Five recent log entries per team leader, checked by their shift manager | Not measured before the course | Day 90: at least 9 of 12 seen doing it in full | Shift managers, at about 30, 60 and 90 days | Find the barrier first (form, time, an unclear rule) before anyone repeats the training |
| Level 3: behaviour B2 | Spoken handover: most urgent first; the receiver reads back the actions | Two handovers watched per team leader, with three checks | Not measured before the course | Day 90: at least 9 of 12 seen doing it in full | Shift managers, on the same visits as B1 | Coach at the handover itself, and check that the handover has protected time |
| Level 2: learning | Team leaders sort events with the three tests and write a note a partner can act on | Card sort score; partner check of three notes | None (new content) | At least 10 of 12 sort 10 or more of 12 cards correctly | Trainer, observer sheet in modules M1 and M2 | Rework the module before the next group and offer a short practice at the follow-up |
| Level 2: learning | Team leaders apply the rules to new cases | Six-item knowledge check at the end of the course | None (new content) | At least 10 of 12 score 5 or 6 | Trainer, 10 minutes at the end of the course | Reteach the weakest item at the follow-up session on 26 Nov 2026 |
| Level 1: reaction | Team leaders know what to do differently and could take part | Three statements from the session feedback form, ticked 1 to 5 | None | At least 10 of 12 tick 4 or 5 for “I know what I am expected to do differently” | Trainer, paper form in the last 5 minutes | Read the open answers and change the part they name |
Evaluation chain
Blank version. Fill it from the top. Add a behaviour row when the course aims at more than two.
| Level | What we want to see | Measure or sign | Baseline | Target and date | Source and owner | If it is not met, we will… |
|---|---|---|---|---|---|---|
| Level 4: result | ||||||
| Level 3: behaviour at work | ||||||
| Level 3: behaviour at work | ||||||
| Level 2: learning | ||||||
| Level 2: learning | ||||||
| Level 1: reaction |
Filled example: evidence table
Status on Friday 12 February 2027: 9 of 12 done, E10 two days overdue.
| # | Level | What we collect | From whom or where | Method | Due | Owner | Status |
|---|---|---|---|---|---|---|---|
| E1 | 4 | Baseline: 40 handover logs from September | Handover logs and the incident list | Count the logs that miss an open issue | Wed 30 Sep 2026 | Quality lead | Done |
| E2 | 1 | Three feedback statements and two open questions | All 12 team leaders | Paper form, last 5 minutes of the course | Thu 12 Nov 2026 | Trainer | Done |
| E3 | 2 | Card sort and note check scores | All 12 team leaders | Observer sheet in modules M1 and M2 | Thu 12 Nov 2026 | Trainer | Done |
| E4 | 2 | Six-item knowledge check | All 12 team leaders | On paper, 10 minutes, marked with the key | Thu 12 Nov 2026 | Trainer | Done |
| E5 | 3 | First-use note: one real handover note in four lines | All 12 team leaders | Brought to the follow-up session | Thu 26 Nov 2026 | Trainer | Done |
| E6 | 3 | Day 30: behaviours B1 and B2 | Each team leader and their shift manager | Five log entries, two handovers watched, 10-minute talk | Fri 11 Dec 2026 | Shift managers | Done |
| E7 | 4 | December sample: 40 logs | Handover logs and the incident list | Same count as E1 | Fri 8 Jan 2027 | Quality lead | Done |
| E8 | 3 | Day 60: behaviours B1 and B2 | Same as E6 | Same as E6 | Mon 11 Jan 2027 | Shift managers | Done |
| E9 | 4 | January sample: 40 logs | Handover logs and the incident list | Same count as E1 | Fri 5 Feb 2027 | Quality lead | Done |
| E10 | 3 | Day 90: behaviours B1 and B2 | Same as E6 | Same as E6 | Wed 10 Feb 2027 | Shift managers | Part done |
| E11 | 4 | February sample: 40 logs | Handover logs and the incident list | Same count as E1 | Fri 5 Mar 2027 | Quality lead | Planned |
| E12 | All | One-page summary for the sponsor | Rows E1 to E11 | Evaluation summary sheet, read together in 30 minutes | Fri 12 Mar 2027 | Trainer | Planned |
Evidence table
Blank version. The Excel register shows the days to each due date and flags overdue items.
| # | Level | What we collect | From whom or where | Method | Due | Owner | Status |
|---|---|---|---|---|---|---|---|
| E1 | |||||||
| E2 | |||||||
| E3 | |||||||
| E4 |
Follow-up conversation at each visit
For the learner’s manager, at about 30, 60 and 90 days. Allow 10 minutes and look at real work, not opinions alone. The learner may decline any question.
| Step | Ask or check |
|---|---|
| 1. Ask the learner | Which part of your handover have you changed since the course? Show me one recent example. |
| 2. Ask the learner | What made it easier or harder to hand over this way? |
| 3. Ask the learner | What would you need to keep doing it? |
| 4. Check B1 | Look at five recent log entries. How many have all four lines? |
| 5. Check B2 | Watch two handovers. Most urgent first? An owner and a deadline for each issue? Did the receiver read back the actions? |
| 6. Name the barrier | Which barrier sits outside the learner’s control, and who can remove it? |
| 7. Record | Write Yes, Partly, No or Not seen for B1 and B2 in the follow-up record, with the barrier and the support agreed, its owner and date. |
Knowledge-check patterns
Seven patterns for a short end-of-course check. The weak items test recall; the stronger items ask learners to do what the job asks.
| Pattern | What the learner does | Use it to check | Weak item | Stronger item |
|---|---|---|---|---|
| Scenario choice | Reads a short work case and picks the best next action from three or four options | Decisions and judgement | “What is the first step of the refund process?” | “The parcel shows as delivered, but the customer says it never came. Which of these three things do you do first?” |
| Spot the gap | Finds what is missing or wrong in a sample of real-looking work | Quality standards and checklists | “List the parts of a good meeting note.” | “Here is a meeting note. Which of the four required parts is missing?” |
| Sort | Puts 6 to 12 items into two to four groups | Rules that classify things | “Define a priority 1 incident.” | “Sort these eight tickets into priority 1, 2 and 3.” |
| Put in order | Arranges steps or items in sequence | Procedures and priorities | “True or false: you isolate the machine before cleaning.” | “Put these five steps of the machine clean in order.” |
| Write it | Produces a short piece of work that someone checks against a list | Notes, replies and requests | “What should a good customer reply include?” | “Write the reply to this complaint. A partner checks it against four points.” |
| Show it | Does the task while someone watches with a checklist | Practical and spoken skills | “Describe how you would give feedback.” | “Give feedback to a colleague in a two-minute practice. The observer ticks three points.” |
| Explain the choice | Gives the reason for an answer in one sentence | Telling a guess from understanding | “Which option is correct?” on its own | “Which option is correct, and which rule tells you so?” |
Before you use a knowledge check
Seven checks for any quiz or knowledge check at work.
| Check | What it means |
|---|---|
| One item, one objective | Each item checks one course objective. Drop an item that checks none. |
| Match the job | The learner decides, sorts, writes or shows, as the job asks. Recalling a definition is rarely the job. |
| Wrong options are real mistakes | Use errors you have seen people make. No joke options and no options that are plainly wrong. |
| No tricks | No double negatives, no “all of the above”, no option that is right on a technicality. |
| Write the key first | Write the correct answer and its reason before the course, and ask a colleague to agree with them. |
| Set the pass mark and what follows | State the mark before the course, and what a learner who misses it gets: another practice, not a penalty. |
| Read the items, not only the people | When most people miss one item, check the item and the teaching before you blame the learners. |
Filled example: knowledge check
Six items for the fictional handover course, each tied to one course objective. Pass mark: 5 of 6.
| # | Objective it checks | Pattern | Case, question and options | Correct answer and why |
|---|---|---|---|---|
| K1 | O1 Sort events with the three tests | Scenario choice | At 05:10 the sensor on dock door 4 fails. The door is closed off until an engineer arrives at 09:00. Is this (a) hand over, (b) log only or (c) leave out? | (a) Hand over. Equipment is out of use and there is an open action for the next shift. |
| K2 | O1 Sort events with the three tests | Sort | Sort four events: a pallet promised to a customer by 08:00 has not left; a spill was cleaned and the aisle reopened; the coffee machine is empty; a damaged rack upright is taped off. | Hand over: the pallet and the rack upright. Log only: the spill. Leave out: the coffee machine. All four must be right. |
| K3 | O2 Write a four-line note | Spot the gap | Read this note: “Door 4 sensor failed 05:10. Door closed off. Engineer coming.” Which of the four lines is missing or incomplete? | The next action has no owner, and there is no deadline: who meets the engineer, and by when? |
| K4 | O2 Write a four-line note | Write it | Rewrite the note for door 4 in four lines. | Correct when a partner can say the next action, its owner and the deadline without asking a question. |
| K5 | O3 Give open issues most urgent first | Put in order | Put three open issues in the order you would hand them over: door 4 out of use until 09:00; the customer pallet due at 08:00; the taped-off rack upright. | Rack upright, customer pallet, door 4. Safety first, then the nearest deadline. |
| K6 | O4 Read back before accepting | Scenario choice | You are taking over. The outgoing team leader says “door 4 is sorted”. Do you (a) accept it, (b) ask what has been done and what is still open, then read back the next action, or (c) check the door yourself later? | (b). “Sorted” is not a state or an action. The receiver asks for the missing fact and reads back the action before accepting. |
Filled example: knowledge-check results
1 is correct and 0 is not. 10 of 12 team leaders scored 5 or 6. Item K3 was the weakest, with 6 of 12 correct: half the group missed the deadline line, so the follow-up session started with it.
| Code | K1 | K2 | K3 | K4 | K5 | K6 | Score (of 6) | 5 or 6? |
|---|---|---|---|---|---|---|---|---|
| TL01 | 1 | 1 | 1 | 1 | 1 | 1 | 6 | Yes |
| TL02 | 1 | 1 | 0 | 1 | 1 | 1 | 5 | Yes |
| TL03 | 1 | 1 | 1 | 1 | 1 | 1 | 6 | Yes |
| TL04 | 1 | 1 | 0 | 1 | 1 | 1 | 5 | Yes |
| TL05 | 1 | 0 | 0 | 1 | 1 | 1 | 4 | No |
| TL06 | 1 | 1 | 1 | 1 | 1 | 1 | 6 | Yes |
| TL07 | 1 | 1 | 0 | 1 | 1 | 1 | 5 | Yes |
| TL08 | 1 | 1 | 1 | 1 | 0 | 1 | 5 | Yes |
| TL09 | 1 | 1 | 1 | 1 | 1 | 1 | 6 | Yes |
| TL10 | 0 | 1 | 0 | 1 | 1 | 1 | 4 | No |
| TL11 | 1 | 1 | 1 | 1 | 1 | 1 | 6 | Yes |
| TL12 | 1 | 1 | 0 | 1 | 1 | 1 | 5 | Yes |
| Correct (of 12) | 11 | 11 | 6 | 12 | 11 | 12 | 10 of 12 |
Filled example: follow-up record
Codes, not names. Yes, Partly, No or Not seen for each behaviour at each visit. Not a rating of any person.
| Code | Shift | Day 30: B1 | Day 30: B2 | Day 60: B1 | Day 60: B2 | Day 90: B1 | Day 90: B2 | Main barrier named | Support agreed, owner and date |
|---|---|---|---|---|---|---|---|---|---|
| TL01 | Early | Yes | Yes | Yes | Yes | Yes | Yes | None named | None needed |
| TL02 | Early | Partly | Yes | Yes | Yes | Yes | Yes | The log form has no deadline field | Quality lead added a deadline field, Tue 15 Dec |
| TL03 | Early | Yes | Partly | Yes | Yes | Yes | Yes | The next shift arrives during the briefing | Shift manager moved the briefing 5 minutes later |
| TL04 | Early | Yes | Yes | Yes | Partly | Yes | Yes | None named | None needed |
| TL05 | Late | No | No | Partly | Partly | Partly | No | Handover is done standing at the dock, in noise | Handover moved to the office desk; shift manager joins two in February |
| TL06 | Late | Yes | Partly | Yes | Yes | Yes | Yes | None named | None needed |
| TL07 | Late | Partly | Yes | Yes | Yes | Yes | Yes | The log form has no deadline field | Same fix as TL02 |
| TL08 | Late | Yes | Yes | Yes | Yes | Yes | Partly | Receiver often arrives late | Late shift start agreed with the receiving team |
| TL09 | Night | Yes | Yes | Yes | Yes | Not seen | Not seen | None named | None needed |
| TL10 | Night | Partly | Partly | Partly | Partly | Not seen | Not seen | Relief cover leaves no overlap with the next shift | Ten-minute overlap added to the night rota from Mon 1 Feb |
| TL11 | Night | Yes | Partly | Yes | Yes | Not seen | Not seen | None named | None needed |
| TL12 | Night | No | No | No | No | Not seen | Not seen | Unsure which events count as open issues | Shift manager went through the three tests with TL12, Wed 13 Jan |
Filled example: result measure
The same count each month: 40 logs checked against the incident list. Target: 10% or less.
| Monthly sample | Logs sampled | Logs that miss an open issue | Share | Target share | Target met? |
|---|---|---|---|---|---|
| September 2026 (baseline) | 40 | 14 | 35% | 10% | No |
| December 2026 | 40 | 9 | 22.5% | 10% | No |
| January 2027 | 40 | 6 | 15% | 10% | No |
| February 2027 | 10% |
Evaluation summary
Blank version, also in the Excel workbook. One row per level, from the result down. Report counts with every percentage and name what else changed at the same time.
| Level | Target | What we saw | Met? | What it means | Decision, owner and date |
|---|---|---|---|---|---|
| Level 4: result | |||||
| Level 3: behaviour | |||||
| Level 2: learning | |||||
| Level 1: reaction |
Filled example: evaluation summary
Interim summary on Friday 12 February 2027, one row per level.
| Level | Target | What we saw | Met? | What it means | Decision, owner and date |
|---|---|---|---|---|---|
| Level 4: result | 4 of 40 or fewer in the February 2027 sample | September 14 of 40 (35%); December 9 of 40 (22.5%); January 6 of 40 (15%) | Not yet | Moving towards the target. The incident list moved to a new system on Mon 4 Jan, which may have changed what is counted, so not all of the fall can be credited to the course. | Wait for the February sample. Quality lead checks ten January logs against the old list by Fri 5 Mar. |
| Level 3: behaviour | At day 90, at least 9 of 12 seen doing B1 and B2 in full | Day 90 so far, in full: B1 7 of the 8 seen; B2 6 of 8. Night shift visits, due Wed 10 Feb, are overdue. | Not yet known | The barriers named were the log form, the rota and where the handover happens, not what people knew. | Night shift visits by Tue 16 Feb (night shift manager). Keep the deadline field and the rota overlap. |
| Level 2: learning | At least 10 of 12 on the card sort and on the knowledge check | Card sort: 11 of 12. Knowledge check: 10 of 12 scored 5 or 6; item K3 (spot the gap) 6 of 12 correct. | Met | People could apply the rules on the day. The deadline line was the weakest point. | Trainer adds a second spot-the-gap practice to module M2 before the next group. |
| Level 1: reaction | At least 10 of 12 tick 4 or 5 for “I know what I am expected to do differently” | 11 of 12 ticked 4 or 5. | Met | The course was clear about what should change. | No change. |
What you'll leave with
In 60 minutes the trainer and the sponsor agree one business measure with a baseline and a target, two or three behaviours someone can see at work, the learning evidence and pass marks, and an evidence table in which every item has a due date and an owner.
Before the session
- For trainers and L&D advisers who plan one course’s evaluation with its sponsor and managers.
- Ask the sponsor for the business measure and its latest figure, or plan who takes a baseline.
- Bring the course objectives and dates. Print the chain and evidence table, or share them online.
- Send plain-text sheets a day early. Allow silent writing and written comments; cameras optional.
How to fill it in with your team
Agree what the evaluation is for
Name the course and its dates. Say: “We are deciding now what would show us that this course worked, so we collect the right evidence at the right time. The evaluation tests the course and the support around it. It does not grade the people who attend.” Ask the sponsor which decision the evaluation should inform, for example whether to run the course again, change it or stop it. Write it at the top of the evaluation chain.
Start with the business result
Fill the Level 4 row first. Ask: “What should be different in the work in three to six months, and how is it counted today?” Write one measure, its baseline with a date and source, and a target with a date. Spend the last three minutes on one question: “What else could move this number at the same time?” List those changes below the chain. If no measure exists, agree who collects a baseline and when.
Name the behaviours at work
Ask: “What would people do differently at work that would move this result?” Give two minutes of silent writing, then agree two or three behaviours that someone can see or check in the work, such as a log entry or a call. For each one, write who can see it, how (a work sample, a short observation or a conversation) and the target at day 90. Ask the managers: “What will you do in the first two weeks to make this easy?” Note their answers in the last column.
Choose the learning evidence
For each behaviour, ask: “What must learners show in the course to be ready for this?” Choose evidence that looks like the work: a sort, a written note, a short practice. Use the knowledge-check patterns table to pick one or two patterns. Agree a pass mark and what a learner who misses it gets, which is another practice and not a penalty. Write both into the Level 2 rows.
Keep the reaction check short
Agree two or three statements for the end of the course. Take them from the session feedback form, and include one about knowing what to do differently. Set a threshold, for example 10 of 12 tick 4 or 5. Say: “Reaction tells us whether the session worked for people on the day. It does not tell us whether the work changed.”
Fill the evidence table and the dates
Turn every row of the chain into one or more lines of the evidence table: what is collected, from whom, by which method, the due date and one owner who has said yes. Put the follow-up visits in the diaries at about 30, 60 and 90 days after the course. Give each result sample the same method as the baseline. Mark anything you cannot collect as dropped, and say why.
Read back and agree the report
Read back the measure and target, the behaviours, the pass marks and the next three due dates with their owners. Agree when the sponsor receives the evaluation summary and that it will say what else changed at the same time. Ask each person for one risk to the plan, aloud, in the chat or on paper, or a pass. Send the chain and the table the same day.
Participant worksheet
Copy this table into a document, or print the worksheet from the bundle. Share only what you choose.
| Item | Detail |
|---|---|
| Evaluation chain | Built backwards: result, behaviour at work, learning, reaction, each with a measure, a baseline, a target and what you will do if it is missed |
| Evidence table | Every item to collect, from whom, by which method, the due date, one owner and the status |
| Knowledge-check planner | Six items that each check one objective, with a pattern, the correct answer and a pass mark |
| Follow-up record | Behaviours seen at about 30, 60 and 90 days, the barrier named and the support agreed |
| Follow-up conversation | Three questions for the learner, two checks of real work and the barrier, at each visit |
| Evaluation summary | For each level: target, what you saw, whether it was met, what it means and the decision |
Debrief questions
- Which business measure did we choose, and what else could move it while the course runs?
- Which behaviour will be hardest to see at work, and who has agreed to look for it?
- Which item in the evidence table has no owner yet, or a date nobody has put in a diary?
- If the result improves but the behaviours do not, what will we tell the sponsor?
Adapt it for your group
- For two people, the trainer and the sponsor, fill the chain together in 40 minutes and send the evidence table to the managers afterwards for their dates and names.
- For a 30-minute version, fill only the result row, one behaviour and its follow-up visits. Add the learning and reaction rows from the course outline after the meeting.
- Online, share one document and fill it while everyone can see it. Collect the silent writing in the chat or in the document, read each row aloud before agreeing it, and keep cameras optional.
- For quiet participants and people working in a second language, give the questions in writing before each step and accept answers on paper or in the chat. Anyone can pass.
- When no business measure exists, or the course is required by a rule, plan the behaviour row in full and use a work sample at about 30 and 90 days as the main evidence. Say plainly in the summary that no result was measured.
After the session
Send the chain and the evidence table on the same day and put every follow-up visit in the diaries. At each due date, update the status in the evidence register and the follow-up record. Before each visit, send the manager the follow-up conversation questions. When the last sample is in, fill the evaluation summary with counts as well as percentages, name what else changed at the same time, and agree one decision with the sponsor.
Questions people ask
What is a training evaluation plan?
A short agreement, made before the course runs, about what evidence will show whether it worked, who collects it and when. This plan starts from the business result the sponsor cares about, then works back to the behaviours at work, the learning in the course and the reaction on the day, so every piece of evidence has a reason to be collected.
Is this based on the Kirkpatrick model?
It uses the four levels of reaction, learning, behaviour and results first described by Donald Kirkpatrick in a series of articles from 1959 and in his 1994 book Evaluating Training Programs. The sheets, questions and wording here are original. Kirkpatrick® is a registered mark of Kirkpatrick Partners, LLC, and this kit is not affiliated with or endorsed by them.
How is this different from the training evaluation form?
The training evaluation form collects participants’ reactions at the end of a session and a self-reported follow-up after two to four weeks. This plan sits above it: it decides before the course which result, behaviours and learning to check, and adds manager visits with work evidence at about 30, 60 and 90 days. Use the form for the reaction row.
Can the plan prove that the training caused the result?
Usually not on its own. Other changes can move the same number, so the summary names them. A baseline taken the same way, behaviours that link the course to the result, and a comparison with a group not yet trained make the case stronger. Report what you saw and what else changed, not a claim of cause.
How do I write a good knowledge check?
Tie each item to one course objective and ask learners to do what the job asks: choose an action in a case, spot the gap in a sample, sort, put in order, write or show. Use real mistakes as wrong options, write the key before the course and set the pass mark in advance. The knowledge-check patterns table gives seven patterns with weak and stronger items.
What is included in the free and editable files?
The open PDF has the blank evaluation chain, evidence table, knowledge-check planner and follow-up record, filled examples of these and of the evaluation summary, the follow-up conversation questions, the knowledge-check patterns and the 60-minute planning meeting. Subscribers also receive the editable Word file, an Excel workbook that counts days to each due date, scores the knowledge check, counts the follow-up visits and compares each sample with the target, and a CSV of the example evidence table.
How to use the files
Print the PDF, edit the Word file in Word or Google Docs, or open the CSV or XLSX in a spreadsheet. Open files need no email.
Can I use this with paying clients?
Yes. Use and adapt this original Superglue kit in your own training, including paid client sessions. Please share this page rather than redistributing the unmodified kit. Read the full usage terms.