# Post-Training Survey — Reading the Answers

> One rule first: numbers tell you where to look; the open answers tell you what happened.
> Use with `post-training-survey-guide.md`. Reading takes 15 minutes for a room of 20.

---

## The 60-second read

Before any averaging, answer three questions in order:

1. **Item 2 mean at 3.5 or higher?** Transfer intent exists. Below 3.0 — the session was theater, and the design (not the delivery) is the suspect.
2. **Q7 commitments countable?** Count how many answers name a concrete action ("update the handover template") versus a vibe ("communicate better"). Concrete answers are the currency you can report.
3. **Q8 obstacles repeating?** Two participants naming the same obstacle is a finding. Three is a redesign order.

---

## Item-by-item interpretation

| Item | Construct | Low (≤ 2.5) means | First action |
|---|---|---|---|
| 1 | Relevance | Content missed the audience's actual work | Rewrite the intake; interview one participant |
| 2 | Transfer intent | Nothing felt usable on Monday | Add one applied exercise per half-day |
| 3 | Pace | Too fast, too slow, or mismatched levels | Split the group or offer two lanes |
| 4 | Facilitation | Questions and dissent were not hosted | Repair the room contract before the next session |
| 5 | Clarity | Exercises ran without a stated purpose | State the "why" in one sentence before each block |
| 6 | Reverse — blocker | **Low is good.** High scores mean participants expect no change | Treat any room-level 6-mean above 3.5 as a design alarm |

The reverse item is the lie detector: polite rooms score everything 4 — including "nothing will change". When item 6 rides high while item 2 rides high too, trust item 6.

---

## Open questions: how to code answers

**Q7 (Monday commitment) → your follow-up list.** Every concrete answer becomes one line in a 30-day follow-up message: "You said you would update the handover template — how did it go?" This is the cheapest Level 3 behavior check in existence: no tooling, one message, honest replies.

**Q8 (obstacles) → your redesign backlog.** Sort answers into three buckets: environment (room, time, tech), design (exercises, sequencing), and outside-the-room (workload, manager support). You control the first two; say so honestly when reporting the third.

Discard answers that only praise. Praise is pleasant and information-free.

---

## Reporting to a sponsor or HR

Report four numbers and one sentence — never the full sheet:

1. Response rate (your credibility: 90%+ in-room beats 30% by email)
2. Item 2 mean (transfer intent)
3. The lowest-scoring item, by name, with its number
4. Count of concrete Monday commitments from Q7
5. One verbatim obstacle from Q8, quoted with permission

Never report: individual sheets, per-person rankings, or the overall average alone. An average without the lowest item is how training reports lose their reputation — sponsors have seen too many 4.6s that changed nothing.

---

## 3 known failure modes of reading survey data

1. **Cherry-picked praise.** Quoting only the warm Q7 answers turns the survey into marketing. Rescue: quote one obstacle for every compliment — the contrast is what sponsors trust.
2. **One angry sheet overrides the distribution.** A single hostile response is data about one experience, not the room. Rescue: report distributions, and read extreme sheets last, not first.
3. **Survey scores used as facilitator KPI.** The moment scores become personal performance metrics, the instrument is dead — answers turn diplomatic and worthless. Rescue: room-level reporting only, in writing, on the form itself.