Proposed pilot · Protocol v0.1.0 · 16 October 2026

Thai team CPX assessment pilot plan

A feasibility pilot of baseline, post-training, month 6 and month 12 forms, linking criterion-level KSEC evidence to OBE outcomes to review wording, observation opportunities, burden and scoring before deciding whether to adapt or expand use.

Download plan and recording guide (MD)Choose recording forms

Analyze stakeholders through HADR tasks, proposed CLO/PLO and Bloom levels

This project proposal has no pilot results, endorsement or validity/reliability evidence. Its primary purpose is feasibility, not proving training efficacy, certifying readiness or ranking teams. The 6/12-month schedule and 0–3 rubric are proposals, not research-validated standards. The website provides blank forms and collects no team, personal or incident data.

Lock the plan before T0

Scope and team unit

Purposefully include Thai teams with contexts relevant to the question, such as agency, locality, remit, experience and resources. Record invitation, participation and non-participation reasons; do not assume national representativeness. A methodologist should help determine team sample size, desired precision and resource limits before collection; this plan supplies no fixed sample size.

The team is the follow-up unit. Use coded team and member/role records, tracking members continuing from T0/current roster size and role changes at each wave. Prespecify how changed team composition will be handled. The same organization name does not establish the same team, and eight criteria are not eight independent team observations.

Roles, authority and permissions

Name a coordinator for scheduling/document versions, controllers for injects and prompting logs, independent raters A/B, a disagreement reviewer, a data steward and a methodologist. The responsible local official checks applicable plans/orders with edition, date, clause and proposal, coordination, approval and publication roles. Record overlapping roles and conflicts of interest.

Before starting, have the responsible institution determine whether the activity is service improvement or research and what ethics review, consent and data permissions apply. This plan neither grants permission nor establishes legal compliance. Specify access, retention/deletion, reuse and reporting in authorized institutional systems. Use codes and minimum necessary data, separating learning from individual punishment.

Specify stop authority, concern-reporting channels and the process for handling concerns and authorizing restart. Mark materials exercise only and do not task real agencies. Stop for safety or data concerns or a real emergency; report to the responsible official without allowing assessment to disrupt response.

Approve the protocol before T0

Obtain responsible-owner approval of the protocol, forms, framework version, review criteria and analysis plan before T0. Specify the follow-up reference date, day-counting rule, acceptable M6/M12 windows, out-of-window reporting, wave comparisons, denominators, exclusion reasons and missing-data handling. Log later amendments with reasons rather than silently changing rules after seeing results.

Check all eight criterion–KSEC–OBE–inject–artifact–observation-window links against the canonical framework and curriculum, including Thai/English comprehension and accessibility. C concerns observed contextual conduct, not personality. Prepare fresh A–D scenarios from one blueprint and record opportunity, difficulty, time, resources and prompting. A shared blueprint does not establish scenario equivalence.

Raters use the calibration kit and recheck with new unseen cases, recording prior key exposure. Before data exist, a methodologist specifies ordinal agreement analysis, uncertainty estimates and required team numbers. Do not invent reliability values or unsupported passing thresholds.

Record, lock and reconcile

Use separate A/B forms at every wave. Record evidence IDs, time, behavior, opportunities, prompting, rating and rationale without seeing the other rater's score. Mask prior scores/phase where feasible and disclose limits. Timestamp and lock originals before discussion. The reviewer adds resolution, reasons and remaining dissent without overwriting original ratings; rank or averaging does not resolve differences.

Use 0 when an observable opportunity existed but the behavior was not demonstrated; reasoned NO for unobserved behavior, insufficient evidence or no opportunity; and justified NA only for role/scope inapplicability declared before the wave. Blank means unentered. Keep all three separate from zero and distinguish a wholly missed wave; never carry an earlier score forward into a missed wave.

Criterion analysis and separate incident evidence

Report paired 0–3 level transitions by criterion for prespecified wave comparisons, restricted to sufficiently comparable teams, opportunities and scenarios, with available/expected pairs and ineligibility reasons. Show independent scores and pre-consensus differences, separating NO/NA/blanks. Do not total scores, average ordinal levels, calculate percentage retention or infer causality. At every wave show invited, enrolled, scheduled, observed and missing teams with reasons.

Keep repeat CPX ratings, between-wave corrective actions, participant reports and incident forms in separate streams. Record only naturally occurring, authorized incidents with training-to-incident interval, continuing members, use opportunities and context. Separate self-report from contemporaneous documents a reviewer can verify; record use, non-use, failure, contradictory evidence and unverifiable claims. No incident is not zero, and incident forms do not replace M6/M12.

Make a feasibility decision

Review after each wave and synthesize after M12. Use the qualitative signals below with evidence and limitations, not invented readiness cutoffs. Record the rationale for holding, adapting, repiloting or expanding, including dissent, owner, conditions and review date. Expansion requires a documented rationale and authorization; improved ratings do not trigger it automatically.

One team, four observation waves

Prespecify scheduling windows, record actual dates and deviations, and never fill a missed wave with an earlier score.

T0
Before training
Before the teaching, training or improvement intervention being evaluated; record date, time and time zone.
- Confirm the approved protocol, rubric version, data permissions, KSEC/OBE blueprint, roles and team register. Collect T0 before teaching or the evaluated training intervention, using unseen scenario A and a standard briefing.
- Record date/time/zone, roster and roles, prior experience, opportunities and prompting. A/B lock ratings/evidence before review. Log deviations; if a true T0 is absent, do not reconstruct it retrospectively.
Deliverables
- Locked T0 A/B forms, artifact/opportunity set, eight-criterion profile, baseline team register and deviation log.
Wave context form (CSV)
T1
Immediately after training
Begin immediately after the specified training ends. Record training end, assessment start/end and actual elapsed interval; disclose any delay.
- Assess after the specified training and before assessment feedback/AAR, using fresh scenario B. Record training end, assessment start/end, actual interval, training content, fatigue and deviations. Simulated feedback required for TC-C2 is a standard task inject, not disclosure of assessment results.
- Lock A/B and evidence before feedback, review A/B scenario comparability and roster changes, then conduct the AAR and maintain corrective actions separately from ratings.
Deliverables
- T1 A/B forms and review trail, criterion-level T0–T1 paired table with denominators, actual timing/deviation log and separate AAR register.
Wave context form (CSV)
M6
Month 6 follow-up
Target six months after training completion: a project proposal. Prespecify a scheduling window and report actual elapsed days.
- Schedule within the month 6 window declared before T0. Record the assessment date and actual days from the reference date with in/out-of-window status. Use fresh scenario C and review opportunities and difficulty without assuming equivalence to A/B.
- Record continuing/current members, new roles, refreshers, other training, task exposure, incidents and plan/resource changes. Lock A/B before discussion. For a missed wave record reason and follow-up status, without reusing an earlier score.
Deliverables
- M6 A/B forms, artifacts, prespecified paired-wave tables with denominators/limits, context-change register and observed/missing team register.
Wave context form (CSV)
M12
Month 12 follow-up
Target twelve months after training completion: a project proposal without demonstrated Thai team retention at this interval. Record planned window and actual date.
- Use the prespecified month 12 window and record actual days as at M6. Use fresh scenario D and recheck opportunities, difficulty, time, prompting and team context. Do not attribute performance solely to initial-training retention when intervening training or other factors occurred.
- Collect and lock A/B, review prespecified wave pairs, attrition and dissent, and synthesize feasibility with raters, teams, the data steward and methodologist. Keep incident and corrective-action evidence separate from ratings.
Deliverables
- M12 A/B forms and evidence, across-wave criterion report with missingness/denominators/uncertainty, decision record and version-revision or repilot plan.
Wave context form (CSV)

Blank forms for institutional data collection

This site receives or stores no study data. Download copies into authorized institutional storage. Lock A/B forms separately before comparison; use codes and keep identity registers separate.

Use one set per team. Copy transfer per incident/task and decisions per reviewed wave/criterion. Actual data fields are blank; criterion/version codes are prepared. Blank is not zero. P/R tasks train raters; they are not T0/T1/M6/M12 team outcomes.