HCD.14 — Revise the Human Capability-Development Arrangement from Evidence
Type: Human Capability Development practitioner method pattern Status: Stable
HCD.14:1 - Problem Frame
Use this when an observation of performance, unfamiliar transfer, delayed performance or support dependence changes an assumption behind a person’s development arrangement. Begin with the affected observation and ask which part of the arrangement it calls into question.
The first useful result is a bounded revision hypothesis: a specific change to the arrangement, the contribution it is expected to improve, what remains protected, and the next observation that could retain, adapt or reject the change. A compatible result from HCD.11, HCD.12 or HCD.13 can be sufficient; the three are not obligatory stages.
An arrangement includes the tasks, practice, feedback, sources, tools, provider and other conditions through which development is attempted. The Method revises the smallest justified part of that arrangement. It does not treat every disappointing result as a reason to rebuild a course or select a new life or career direction.
Do not proceed from a score with no recoverable contribution or observation. Obtain the missing evidence first. If the live question is clinical treatment, professional correctness, safety, employment action or an unavailable domain Method, return it to the responsible specialist or owner. A new source edition alone does not establish deterioration in the person.
HCD.14:2 - Problem
A development plan embodies hypotheses: this task will elicit the needed contribution, this feedback will help, this support will remain available, and this practice will prepare the person for later work. Evidence can overturn one hypothesis while leaving the others usable.
Two failures are common. One repeats the same exercise more often without identifying the discrepancy. The other redesigns everything, losing supported contributions and spending resources without evidence that the whole arrangement was at fault. Both can conceal a defective assessment, a cue that supplied the answer, or an unavailable source behind a claim that the person needs more training.
HCD.14:3 - Forces
| Force | Working tension |
|---|---|
| Action and causal uncertainty | Make a small testable improvement when useful without pretending that one miss establishes its cause. |
| Local repair and shared causes | Preserve working components while enlarging the change when evidence shows that several failures share a cause. |
| Learning and assessment | Feedback can be useful during practice but must not silently supply the contribution later being assessed. |
| Professional Method and local tailoring | Adapt a qualified development Method to the case without inventing expertise from a pattern label. |
| Improvement and protection | A gain in speed or score is unacceptable if it damages a required contribution, access or resource limit. |
HCD.14:4 - Solution
HCD.14:4.1 - Recover the assumption that the evidence changed
Name the person, intended later Work, human target, affected observation and receiving decision. Identify the arrangement component expected to help and the assumption now in doubt.
Use a concrete discrepancy: “the person applies the capacity rule after a cue but repeats the old promise when available time changes,” not “transfer is weak.” Preserve successful contributions and the conditions under which they were observed.
If the evidence and target concern different people, work families, configurations or periods, qualify their relation before using the result. A cohort average cannot by itself identify this person’s limiting contribution.
HCD.14:4.2 - Compare only causes that change the revision
Ask what else could explain the discrepancy and what different action each explanation would imply.
| Live explanation | Different useful next move |
|---|---|
| The target or later-Work demand changed | Return the changed demand to its owner; preserve the old claim within its former conditions. |
| The task, criterion or scoring is defective | Repair the assessment and obtain a discriminating observation before revising learning from that result. |
| Required source, tool, time or access was absent | Repair the arrangement with the direct owner; do not infer a human deficit from the missing input. |
| A cue supplied recognition or adaptation | Change the cue/practice relation and test the required contribution before a cue supplies it. |
| The development Method or provider does not fit the target | Obtain a compatible Method or qualified provider result; change that component rather than merely increasing repetitions. |
| A limiting human capability, misconception or behaviour is supported | Use the supported target to select a qualified local development change; retain unresolved rivals. |
A short inspection may settle a known missing source or incorrect criterion. Use HCD.3 when the limiting human target remains unresolved. Use a qualified E.23.CAE differential when access, selection, adaptation, enactment or apparent loss can be distinguished and the distinction changes the next action. Do not add a causal investigation when the current evidence already supports an affordable reversible test and no live rival changes its safe scope.
A small testable revision may follow one well-observed miss. Label it a hypothesis, retain the rival explanation and choose a test that can disconfirm it. Evidence too weak even for that bounded change returns a specific observation to obtain.
HCD.14:4.3 - Specify the positive revised arrangement
State the component to change and the replacement in executable terms. For example, replace a cue-led worked example with two contrasting practice cases and feedback on the person’s own first explanation, then test a fresh uncued variant. “Improve transfer” or “use HCD.5” is not a revision.
Select the professional development Method from an applicable qualified source or a compatible direct provider result. Check that its target, learner conditions, prerequisites, support and limits fit this use. If the Method or provider competence is missing, return that exact input.
State what must not change: supported contributions, required checks, accessibility arrangements, source access, permitted tools, time and cost limits, or specialist responsibilities. A change to the target itself needs the target owner’s decision.
HCD.14:4.4 - Pair the change with a test and stopping rule
Name the expected observable difference and the fresh task or other observation that could show it. Include a meaningful failure condition and an observation that would point to a rival cause.
The test should elicit the affected contribution without reproducing the cue the person is supposed to supply. Preserve domain criteria and compare first performance separately from feedback-assisted correction. Check protected contributions as well as the intended gain.
State acceptable learner and practitioner cost, resources, timing and any delayed observation actually needed for the claim. Do not add a retention visit merely to complete a sequence. When delay matters, HCD.13 supplies that comparison under its own resources and evidence conditions.
Specify what to retain, adapt, abandon or return after the observation. A local before/after comparison can guide the next reversible revision; an estimate of causal effectiveness needs a suitable causal design and evidence beyond this comparison.
HCD.14:4.5 - Distinguish proposal, authorization and performed change
Return the revision as a proposal when the practitioner lacks the mandate or resources to carry it out. Name the actual owner of the resource, provider, target or authorization gap and what that owner must supply.
Within an obtaining mandate, the responsible practitioner can make the bounded change and then observe its result. Keep the selected proposal, authorized arrangement, performed development work and observed outcome distinct. A plan and a changed document do not establish that a person practised or learned.
Stop if the change would breach a protected condition or exceed the mandate. Use a safe restoration of the prior working component when needed; do not continue simply to obtain a cleaner comparison.
HCD.14:4.6 - Update the hypothesis from the new observation
Retain the change provisionally when the intended contribution improves under the test conditions without unacceptable loss. Adapt it when the result narrows the useful conditions or identifies a more local problem. Abandon or return it when the decisive prediction fails, a protected contribution worsens or the professional basis no longer fits.
Preserve unchanged claims and surviving rivals. Reopen only the assumptions affected by new evidence. Enlarge the revision when a changed demand or several independent observations show that the problem is not local; this Method does not require keeping an inadequate component merely because the first proposal was small.
HCD.14:5 - Archetypal Grounding
HCD.14:5.1 - A local revision at the transfer return
Learner-L17 and the evidence are fictional. The case begins at a transfer return, before any delayed test is required. It does not report an actual programme change or learner outcome.
The earlier evidence supports calculation and configuration checking. In a new case, L17 correctly uses a complete HW3/FW8/SCHED2 report and proposes ten changes within 24 available hours. Required checks take two hours per change; preparation takes one hour per homogeneous batch of at most four. Ten changes need 23 hours, twelve need 27, so a ten-change proposal leaves two of twelve new weekly changes in the queue. The release manager retains release/use authority.
When available time changes to 22 hours, L17 initially keeps the ten-change promise and corrects it only after feedback. Nine changes need 21 hours. The calculation and configuration evidence are not erased by this adaptation failure.
The prior practice exercise used a teacher question to point out the changed resource condition. The live hypothesis is that practice did not sufficiently require L17 to detect and adapt to that change before the cue. A single miss does not settle the cause: unclear task wording remains a rival, and general resource reasoning may also need investigation.
For this constructed case, the programme owner has supplied a compatible case-practice Method and a qualified practitioner. The Method asks the learner to attempt a case, explain the consequential choice, compare it with a contrasting case, receive criterion-based feedback, and attempt fresh material. Those are stipulated professional inputs to the example.
The proposed replacement is concrete:
| Element | Revision hypothesis and protection |
|---|---|
| Component changed | Replace only the cue-led resource exercise. Leave the supported financial calculation and configuration practice in place. |
| First five minutes | Recover L17’s explanation of the earlier promise and inspect whether the task clearly presented the changed time. If it did not, repair the task before interpreting the miss as a practice problem. |
| Twenty minutes of practice | Use contrasting cases requiring a supported ten-change proposal at 24 hours and refusal of that same promise at 22 hours, with a feasible nine-change proposal. Obtain L17’s first explanation before feedback names the decisive condition. |
| Ten-minute fresh test | Give unrehearsed 20-hour and 28-hour variants with complete compatible reports and twelve new weekly changes. Do not name the resource Method or point to the changed number. |
| Expected contribution | At 20 hours, eight changes require 18 hours and nine require 21: do not promise nine. At 28 hours, twelve require 27: a proposal to meet those twelve is supported. Preserve required checks, source fit and any existing-queue limitation. |
| Protected conditions | Same source and calculator access, ordinary permitted AI assistance, accessibility arrangements, domain criteria and separate release authority. Observe L17’s first recognition and explanation in the source and assistant traces; a teacher or AI cue that supplies recognition of the changed condition counts as help. |
| Cost and status | Forty learner minutes including a final five-minute result discussion, plus fifteen practitioner preparation minutes, within the example’s proposed replacement allocation. The return is a proposal until its responsible owner authorizes that use of time. |
The fresh task tests both justified restraint and a justified positive action. It does not award success for repeating the earlier refusal.
HCD.14:5.2 - The next observation decides the smallest update
No result of the proposed revision is assumed. The test has the following conditional returns:
| New observation | Update to the hypothesis |
|---|---|
| Both fresh uncued first attempts adapt the promise correctly and preserve configuration checking and calculation | Retain the local change provisionally for these tested variants; do not claim a general transfer effect. |
| Correct answers appear only after a teacher or AI cue names the changed resource | Independent adaptation remains unproven; inspect cue dependence and the supported human target before adding more repetitions. |
| The time value is ambiguous or a required source is unavailable | Repair that assessment or access component; the affected attempt cannot judge the intended learning contribution. |
| Adaptation improves but configuration checking or required checks deteriorate | Revise or abandon the change: the intended gain does not compensate for the protected loss. |
| Further qualified observations show several tasks depend on the same missing prerequisite | Enlarge the target or arrangement question through the appropriate diagnostic and Method owner. |
A later HCD.13 result may update the hypothesis if delay or support dependence becomes relevant. It is neither a prerequisite for this first revision nor proof that the revision caused the later result.
HCD.14:6 - Bias-Annotation
A familiar teaching technique can become the default explanation for every failure. Compare the assessment, support and target rivals before selecting it. Conversely, uncertainty about the final cause should not prevent a clearly bounded, affordable and reversible test.
Protect the learner from a sequence of ever-changing exercises justified only by a rising score. Ask what got worse: independent selection, source access, correctness, workload, confidence in a justified action or a previously supported contribution.
HCD.14:7 - Conformance Checklist
- The affected observation changes an identifiable development assumption for the same person and work.
- Successful contributions and decision-changing rival causes remain visible.
- The result names a concrete component and an executable replacement, not only another pattern.
- A compatible qualified development Method or exact missing professional input is identified.
- The expected observation, fresh test, cost, protected conditions and stopping rule are explicit.
- Proposal, authorization, performed work and outcome remain distinct.
- Retain, adapt, abandon and return conditions affect the smallest justified claim or component.
- No whole-programme causal effect or new development direction is inferred from a local comparison.
HCD.14:8 - Common Anti-Patterns and How to Avoid Them
| Anti-pattern | Better move |
|---|---|
| “The score fell, so prescribe more practice.” | Recover the contribution and observation, then compare the causes that change the revision. |
| “One miss proves the training Method failed.” | State a bounded hypothesis and a discriminating fresh test. |
| “Replace the whole course.” | Preserve supported components and enlarge the change only when evidence warrants it. |
| “The new plan means the person has developed.” | Separate the proposal, performed work and observed outcome. |
| “The final score rose, so the revision worked.” | Check independent contribution and protected conditions; qualify any causal claim. |
HCD.14:9 - Consequences
The practitioner returns a usable improvement proposal without pretending that the evidence settled every cause. Local revisions preserve successful learning and make the cost and next observation inspectable.
This approach can require saying that the missing input belongs to the assessor, source owner or provider rather than to the learner. It also accepts provisional results: a useful change may be retained locally while its broader transfer or retention remains untested.
HCD.14:10 - Rationale
An arrangement is revisable because it embodies claims about how a person will develop for later work. The right unit of repair is the smallest assumption changed by evidence, not automatically the whole programme or the person. An explicit test makes a revision useful even while a causal rival survives, provided the claim, cost and protected conditions remain appropriately bounded.
HCD.14:11 - SoTA-Echoing
Working question: how should performance evidence change development when the cause may lie in the task, feedback, support or learning?
Selected line — adopt and adapt. The Kraiger and Ford workplace-instruction review separates instructional design, learner activity, learning outcomes and work transfer. HCD.14 uses those distinctions to locate the changed assumption in §4.1–§4.3, instead of treating every outcome as a verdict on the entire programme.
Serious alternative: repeat or redesign the whole course after a poor result. That is appropriate when the demand or evidence shows a broad mismatch, but wastes supported contributions when only one cue-led exercise is implicated. At comparable decision effort, the selected line returns an executable component change, a test and protected results. Its deliberate limit is that a small trial supports a local update, not general effectiveness.
The Bastani and colleagues study makes assistance design a credible rival in its school-mathematics setting; it supplies no adult-provider prescription. The Yang and colleagues testing synthesis makes testing and feedback conditions relevant to learning. These limits change §4.4 and §5.2: first and corrected attempts stay separate, and a later improvement is not automatically a causal effect.
Reopen the revision if a protected contribution worsens, a rival explanation changes the useful repair, the supplied professional Method does not fit, or broader observations overturn the local scope. The positive proposal in §5.1 remains a hypothesis until its authorized test supplies evidence.
HCD.14:12 - Relations
HCD.11, HCD.12 and HCD.13 can each supply compatible evidence; none is a compulsory predecessor to the others. HCD.3 supplies a qualified limiting target or non-training return when diagnosis remains unresolved. HCD.4 can use changed contribution evidence in its existing profile comparison.
HCD.5, HCD.6, HCD.9 and HCD.10 concern development Methods, practice, providers or support under their own scope. Use the applicable qualified professional contribution, or return the exact missing Method or provider input. Do not reconstruct those Methods from this revision pattern.
E.23.CAE supplies a useful differential when apparent loss, access or expression is at issue. C.11 governs an actual choice among existing alternatives when its chooser and comparison conditions obtain; the target, resource and programme owners retain their decisions. A causal-effect claim requires an applicable causal evaluation, including C.28 when its question is live.
The Learning Product owner changes the concrete programme or instrument. The person’s development direction, commitment, clinical care and employment remain separate questions; a new HCD finding can inform them but does not decide them.
HCD.14:End