HCD.19 - Evaluate Instructional Material through Representative Reader Work and Diagnose Its Repair
Type: Human Capability Development practitioner material-evaluation and repair-diagnosis pattern Status: Stable
HCD.19:0 - Use This When
Use this pattern with a compatible specification to establish what instructional material supplies for a promised use. Examine its contribution, any limiting defect and, where needed, the repair or evidence that would change the conclusion.
Start with the promised task, its subject criterion, the material and the assistance available to the audience. Use inspection or existing compatible reader work where it answers the question. If the claim depends on a reader’s response, examine what the reader did and which passages, examples or help could account for it. For a new test of independent recognition, preserve the first answer before giving an answer-bearing cue.
The first useful result is a qualified material judgement: a supported contribution, a limiting defect and its reach, or a specific evidence or criterion gap. Use that result for the requested assessment, a use decision or an improvement under E.23. A correct answer can coexist with a defective explanation if the reader repairs it from prior knowledge.
Use HCD.18 when a suitable specification must be selected, constructed or repaired. For a person’s performance, transfer or retention, use the relevant method in HCD.11–HCD.13. If a subject claim or exercise key is disputed, consult the applicable domain grounds. Preserve any available reader answer and state which conclusions the dispute leaves open before attributing a disagreement to comprehension.
HCD.19:0.1 - Working Terms
| Term | Meaning in this pattern |
|---|---|
| Reader task | The contribution requested from a reader to expose support for a promised use. It may be a decision, calculation, explanation, repair, or other domain-valid action. |
| First answer | The response obtained before later feedback or a cue supplies the distinction whose independent recognition matters. The ordinary material and declared tools may already be available. |
| Supported continuation | A later response made with the recorded help. Its usefulness is judged under that support arrangement. |
| Material contribution | The distinctions, grounds, operations, examples and usable ways to reach the needed source, help or next task that the material supplies. |
| Limiting defect | A supported material or support deficiency that prevents the declared use or requires repair, with its effect traced to the affected tasks. |
| Repair reach | The connected explanation, example, task, criterion, representation, or return that must change for the defect to be corrected in use. |
HCD.19:1 - Problem Frame
An editor receives two apparently conflicting results. A prepared reader solves an interrupted-task exercise correctly, while the handout defines task effort as start-to-handoff time. The reader explains that the active/elapsed distinction came from prior knowledge. After a teacher question and a corrected explanation, the answer remains correct.
Inspection identifies the handout’s false relation and the corrected relation in the revision. The reader already supplied the distinction, so the unchanged answer leaves a learning gain unestablished. A task with two people can test whether the revised material supports the intended population denominator. Evidence of later retention would require an appropriate interval.
A different difficulty arises in a long course. A reader may understand each local definition but fail to connect an early distinction to a later choice. Evaluating isolated excerpts will miss that edge. Conversely, a wrong figure return can be repaired locally while a reported broader difficulty remains unresolved.
Use the task, its criterion, the relevant material and any reader work needed for the inference to establish what the material contributes. Where a defect is supported, locate the smallest complete repair. A change to the material, a learner prerequisite, teacher support and an evaluation criterion have different effects.
HCD.19:2 - Problem
How can an evaluator obtain and interpret representative reader work so that the resulting material judgement distinguishes defects, prior knowledge, assistance, and missing evidence, and identifies any needed repair and how to test its effect?
HCD.19:3 - Forces
- A natural open-book task can be representative even when a worked example is nearby; answer proximity still limits what the result establishes.
- First recognition and successful correction are both useful, but a later cue can erase the opportunity to observe the first.
- Expert inspection can reveal a subject defect cheaply, while audience-dependent adequacy requires evidence from the relevant reading conditions.
- Whole-route claims need connected reading; a local repair should reuse unaffected evidence.
- A teacher can legitimately supply a promised contribution. Requiring all help to reside in the material changes the product being evaluated.
- Useful difficulty may prepare later instruction, whereas unsupported difficulty can leave the task impossible. The next instructional move distinguishes them.
HCD.19:4 - Solution
Evaluate the material against a compatible specification. Qualified source inspection or existing observations may establish the requested finding. Select further reader work when it could supply needed evidence, the necessary participants and provider capability, time, access and protections are available, and its attainable contribution warrants the design, preparation, reader and provider work, interpretation, delay and displaced work. The instructional-material evaluation profile supplies common property meanings. The specification selects the properties, scales and evidence. Identify any audience, experienced-burden or learning claim that the available evidence leaves open.
HCD.19:4.1 - Fix the material, criterion, and ordinary support before the attempt
Identify the material edition, task, audience prerequisites, and intended question or decision. Check that these conditions match the specification. Preserve a current material copy or precise edition reference when later repair would otherwise make the observation unrecoverable.
Derive the task from the promised work. For a new observation, establish the domain criterion, valid alternatives and any consequential rival response before seeing the answer. To interpret an existing response, establish the applicable criterion independently of that response. When independent selection matters, leave the choice of rule to the reader. A prompt naming the distinction can test its application.
For a new reader task, provide the material and declared ordinary resources. For an existing observation, establish what the reader could use. Keep the evaluator’s answer key, intended defect, unpublished rationale, and later condition unavailable when they would supply the contribution being tested. A public worked example legitimately available in reference use is part of the material’s support; record that closeness instead of treating it as a test violation.
For a new observation involving people, explain its purpose and conditions to the participants. Do not require them to experience a foreseeable harmful consequence just to demonstrate a source defect. Subject inspection and a safe task can often settle the repair before a higher-risk use.
If the task or criterion is disputed, preserve any existing response. Use applicable domain evidence to settle the issue where possible or identify the conclusion it leaves open. The exercise key establishes what it accepts. Its correctness is a separate question.
HCD.19:4.2 - Preserve what happened before and after help
When the judgement depends on a first response, preserve it and the reasoning, source use or work trace needed for that inference. Keep errors, omissions, justified uncertainty and valid alternatives visible after correction. If a first answer is unavailable, identify what can still be concluded.
Distinguish the contributions of assistance.
- what the arrangement promised;
- what was available to this reader;
- what the reader received and what they used, including a later cue from the teacher, a peer or an AI system.
When attribution matters, check what the material contains and what the response shows. A reader’s account can distinguish using the material from drawing on prior knowledge, although a familiar inference may go unnoticed. Seek that account when it would supply missing evidence. A causal claim requires grounds for the particular attribution.
For a selected FPF C.2.8 (another source edition) structural-recovery question, keep the correctly recovered relations and their attribution at these conditions. Familiar knowledge may enable decoding; correcting false or absent instruction from memory does not give the material credit for that correction. A response can correctly report what the material asserts without establishing that assertion’s subject truth. Compare claims and forms separately when a revised material changes content. One response is a bounded recovery observation, not a maximum extractable amount or a learning result.
For a planned feedback episode, preserve the first response and provide the declared feedback. Record what help was given, when it was given relative to the response, and what followed. For a report on completed work, use the available evidence of those events. Distinguish an explanation that supplies the answer from a prompt to inspect an available criterion; both can be useful under their respective conditions.
Select reader and task evidence for the property, audience and consequences in question. A cold intended reader lacks the author’s private context but brings their own expertise. A prepared agent’s performance can establish its reading under those conditions. Claims about novice human performance need evidence about that audience.
HCD.19:4.3 - Read the material at the reach of the claim
Inspect the definitions, examples, representations, criteria, links and prerequisite relations on which the task relies. Follow a dependency until you find the required contribution or establish what is missing or inaccessible. A heading can locate the passage; read the content needed at its point of use.
For a judgement about a whole route, read the entire declared route in its intended sequence and check which passages the audience can reach where each contribution is needed. If observed readers took another route, identify it when it affects the claim. Follow consequential dependencies to their points of use. When several readers examine parts, the integrator needs connected evidence for the transitions and whole route.
At a narrative transition, use NSTD.8’s attachment question: what question is current, what does this block add, why is it needed here, and where is its contribution used next? Inspect both endpoints. At the next use, distinguish structure already introduced or available through legitimate prerequisites from the genuinely new relation the reader must construct.
A later explanation can support the present task when a precise, usable return makes it available now. Without that access, locate the missing support where the reader first needs it. The repair may be one contrast, a worked step, a map, a recap, or a changed order.
For an existing product, current content and reader-use evidence can support a current-quality diagnosis without original manufacturing records. If source fidelity or construction history is part of the promise, use the separate evidence branches of NSTD.6. A plausible reconstruction of how a book could have been written is not an observation of its author’s actions.
HCD.19:4.4 - Judge the contribution before assigning a material value
Where there is a reader response, compare it with the qualified criterion. Distinguish a correct result, an error, an unfinished attempt, an unobserved contribution, and a task or criterion defect.
Trace the material’s part in the outcome. Different explanations can warrant different checks or repairs.
| Observation | What to distinguish | Discriminating next move |
|---|---|---|
| The reader cannot justify the next operation | Missing explanation, missing prerequisite, or a question the material never promised to answer | Inspect the promise and the preceding relation. Use existing evidence about help or try the least additional declared aid capable of distinguishing these explanations, when that inquiry is warranted. Preserve what the reader answered before it. |
| A correct-looking example supports a wrong rule | A non-discriminating example versus a false or absent subject relation | Use a case on which the intended and rival rules produce different results; inspect both rules explicitly. |
| The key rejects a valid response | Criterion defect versus reader error | Check the domain criterion and preserve the response. If the key rejects a valid alternative, correct the key where a repair is requested. If the criterion is still disputed, state the limit. |
| Separated concepts can be repeated but not combined | Missing relation, lost attachment, or unavailable prefix support | Inspect the two endpoints and the exact point of use; identify the missing relation or access and repair it where required. |
| Success begins after the teacher names the decisive condition | Effective supported continuation versus independent recognition | Retain the helped result; use applicable evidence from an uncued response, or a fresh task or reader if warranted, for a claim about independent recognition. |
| An expert succeeds despite a false explanation | Prior-knowledge rescue versus support supplied by the material | Preserve the source defect and withhold untested adequacy for less-prepared readers. That stronger use needs qualified audience evidence; select its collection only for an obtainable, worthwhile contribution. |
| A removed sentence makes no difference | Another public passage supplies the contribution, or the perturbation was irrelevant | Record the successful recovery. Revise the defect hypothesis rather than redefine success as failure. |
| A prepared agent succeeds while a human reports difficulty | Different knowledge, task, help, or experienced burden | Keep the human report and narrow the untested human claim now. Preserve the supported material diagnosis. Select a timed observation of the relevant reader and segment only when obtainable under the declared conditions and worth its full burden for the requested finding or decision. |
Interpret success or failure against the criterion, the material, prerequisites and available help. These can explain the same response in different ways.
Apply the specification’s anchors to the material condition the evidence establishes. In a full evaluation, return each selected property with its value or the reason it is undetermined. A narrower diagnostic identifies the task and properties it covers. For adjacent values, check the meaning of the anchors and the reach of any repair. Report unresolved disagreement without averaging ordinal ranks.
HCD.19:4.5 - Diagnose the smallest complete repair
When you find a defect, state what it prevents or misrepresents. “The report rule includes incident time in effort” identifies a repairable relation; “unclear” does not. Locate the explanation, task, criterion, representation, or support arrangement in which that relation changes use.
Trace where the affected relation is used. A central distinction used in a definition, table, interruption example, and later resource comparison needs a connected repair. A wrong figure address can require only the return and its receiving use. Several affected words are not necessarily several defects; one wrong premise can control them all.
For a repair recommendation, propose a replacement and identify the reader result it should enable. Preserve subject truth, valid alternatives, meaningful expressive details, accessible representation and the help promised by the course. Identify any cost or trade-off that could make another repair preferable.
If repair selection or iteration is part of the task, use E.23 with the supported defect, its dependent uses, the useful contributions to preserve and a test that could distinguish a successful repair. A diagnosis can also be the requested result.
If promised teaching support is unavailable, identify the missing service or condition. When the assignment calls for a remedy, return that need to whoever is responsible for the arrangement, if known, and identify who can supply it. State any unresolved responsibility or supply. If a prerequisite is missing, compare accessible help, instruction or a revised audience promise when a choice is needed. Carry out a selected change when the assignment requires it and the relevant authorization is in place.
HCD.19:4.6 - Repeat the affected use and stop at the supported conclusion
For a claim about a completed repair, check the changed relation at its task and point of use, using applicable existing evidence where it is sufficient. Reuse observations and values whose premises remain unchanged. Moving an explanation can reopen its former and new attachment and next-use prefix even when its words are identical.
Use a meaningfully changed condition when the question concerns the reach of the repair. Choose a case with a different population, relevant condition, boundary or use so that a plausible overgeneralization would now fail. Preserve the same reader when studying supported continuation; choose a fresh task or reader when prior answer exposure would defeat the needed inference.
A preparatory failed attempt may be retained when subsequent instruction examines and resolves the attempted alternatives. If that resolution is absent, labelling the difficulty “productive” supplies no evidence. Similarly, immediate ease is not the target when the task deliberately asks the learner to construct a new relation with adequate support.
After a repair or feedback episode, report what changed and what stayed the same in the material, the reader’s response and the assistance, to the extent the evidence establishes. Identify any remaining defect or evidence gap. An improved response with help can establish supported performance. Claims about independent performance, unfamiliar transfer, delayed retention or causal learning need their corresponding evidence.
Finish when the evidence answers the requested material question or supports the use or repair decision. State any criterion or evidence gap and which conclusion it limits. A local material finding can stand when an audience, burden or learning claim remains open. To rely on a repaired use, check the affected task. Extend the evaluation to further uses when they depend on what changed.
HCD.19:5 - Archetypal Grounding
HCD.19:5.1 - A correct reader exposes a defective handout
The time-use case in HCD.18 supplies the tasks and criteria. In the recorded clean-context walkthrough, a prepared agent first received M0, which defined task time from first touch to handoff and gave one uninterrupted example. The agent answered using its own arithmetic and time-accounting knowledge.
For one person’s interrupted report, the first response was:
| Requested quantity | First answer |
|---|---|
| Active report work, 09:00–09:30 and 11:30–12:00 | 60 minutes |
| Incident work, 09:30–11:30 | 120 minutes |
| Report elapsed turnaround, 09:00–12:00 | 180 minutes |
| Total active work across both tasks | 180 person-minutes |
| Incident share | 120/180, or two-thirds |
The reader explained that it brought the active/elapsed distinction and person-time meaning from prior knowledge. M0 supplied the intervals and start-to-handoff wording; its uninterrupted example could not distinguish the two quantities.
The supported diagnosis is a material defect despite reader success. The central instruction assigns elapsed turnaround to a question about consumed effort. The subject criterion identifies the defect directly; the observed response shows expert correction of it. Neither observation establishes how a less-prepared human would perform.
The repair M1 states that active effort sums the intervals in which the named person works on the task, while elapsed turnaround runs from first start to handoff. It distinguishes person, task, active intervals, person-time, first start and handoff in the record. It works the interrupted example and gives the rule for comparing available and consumed person-time across people. A worked concurrent-person comparison is still missing.
HCD.19:5.2 - Keep the teacher contribution and test a changed denominator
After the first answer had been preserved, the reader received a teacher’s question about which intervals contained the person’s active work and whether two assigned tasks could consume the same minute of that person’s effort. The reader then received M1 with the numerical answer to the interrupted example.
Its answer remained correct. It identified what M1 now supplied and declined to claim a learning gain: the distinction was already present in its first response, and the worked example supplied the same case’s numbers. The result is a supported walkthrough of the repair.
The later task changed the population. A worked on X from 09:00 to 10:00; B worked on Y from 09:00 to 11:00. Both were available from 09:00 to 11:00, and A did no other work in the remaining hour. The material and calculator remained available; no additional teacher cue was supplied.
The reader returned 1 person-hour for X, 2 for Y, 3 consumed person-hours in total, 4 available person-hours, and a 2-hour clock window. It compared 3 with 4 and identified M1’s rule for the population denominator. The reader applied that rule to numbers the worked example had not supplied.
The changed case distinguishes adding person-hours across two people from limiting the total to one elapsed interval. It supplies bounded evidence that this prepared reader used the repaired rule on a new population condition. It does not show novice learning, delayed retention, experienced burden, or a causal advantage over another explanation.
C.2.8 makes the structural interpretation precise. M0’s correct first answer does not show that M0 expressed the active-effort distinction: the prepared reader supplied the correction. M1 makes corrected relations available, but the teacher cue and same-case answer intervene before its use. The later population task supports recovery with M1 at that changed condition; it does not increase an amount at the initial condition. These facts qualify the structural question without replacing the profile’s material-use and repair values or establishing new learning.
HCD.19:5.3 - Return a complete disposition without inventing a complete success
Time-use S1 calls for a fuller evaluation than the original walkthrough supplied. Check which earlier responses and inspections meet S1’s conditions and identify what is missing.
The following replay evaluates M1 under S1, using source inspection and the preserved observations above. It returns all six selected coordinates. Read its 0–5 values using S1’s anchors for these tasks and audience.
| Coordinate | Value or missingness on this evidence | Grounds, limit, and next action |
|---|---|---|
| Subject validity | Insufficient evidence for S1’s complete subject-validity value | The inspected definitions, interval assignment, interrupted example, and population rule agree with the stipulated meanings. S1 also requires inspection of the offered tasks, key, and correction; that complete arrangement was not inspected for both tasks. Thus 4 is unestablished, while no bounded subject error justifies 3 and no additional checked boundary justifies 5. Retain the supported findings on the inspected passages. Assign the full value only on the inspection S1 requires. |
| Task and dependency coverage | 4, source-inspected for the two S1 task families | M1 supplies interval summation, turnaround, exclusive-person accounting, population aggregation, and the common-unit comparison. It therefore supplies the operations even though one operation lacks a worked example. Missing reader evidence belongs to the explanatory row; the evidence establishes no further consequential coverage advantage for 5. |
| Explanatory sufficiency | Insufficient audience-compatible evidence for a supported-use value | M1 exposes the needed grounds and the prepared reader applies its population rule. That reader already knew the central distinction. The observation cannot qualify the promised introduction for readers possessing only S1’s stated prerequisites. That broader value requires evidence from the intended audience under S1’s conditions. Report the evidence gap and retain the established material findings. |
| Example discrimination | 3, bounded addition required under S1 | The interrupted example distinguishes active effort from turnaround, but M1 has no worked concurrent-person contrast. Its population rule is already present, so adding that contrast does not require rebuilding the central explanation chain. That is 3 rather than 2; the specified example evidence for 4 is incomplete. |
| Task and criterion validity | Insufficient evidence about the complete offered task-and-key arrangement | The original report task is domain-valid and the evaluator’s criteria distinguish the named errors. To assign S1’s value, inspect the tasks and criteria offered for both cases, including what the material or teacher made available to the audience. The evaluator’s private key does not establish that provision. |
| Feedback and continuation support | Insufficient audience-compatible evidence for supported-use adequacy | The teacher question, M1, and later task were supplied. The prepared reader justified retaining its correct answer and continued on the changed task, satisfying S1’s allowed justified-retention branch for this reader. This supports that bounded continuation, but the reader’s prior time-accounting knowledge exceeds S1’s stated audience prerequisites. The stronger audience-compatible value still needs appropriate reader evidence of explained correction or justified retention under the declared support. The current missingness result does not by itself select that inquiry. |
The same missing task-and-key inspection constrains subject validity and task/criterion validity: a correct explanation can coexist with a false offered key. The evaluator’s correct answer supplies neither that component nor evidence of its inspection.
The overall S1 conclusion is repair needed and insufficient evidence for the declared supported use. The 4 for coverage and 3 for example discrimination cannot compensate for the missing evidence. The material needs a worked concurrent-person contrast. Before recommending a change to the task or support, establish what criterion and support the audience can use for both tasks. Supply what the arrangement lacks when the repair is commissioned. A broader audience claim requires compatible reader evidence under that arrangement, with the first attempt and assistance recorded.
This replay gives a worked diagnosis at S1’s declared scope. It uses the compatible observations from the earlier, narrower walkthrough. Evaluating the reliability of S1’s scale requires separate evidence.
HCD.19:5.4 - A long-route defect can remain local
Consider a designed course route that introduces a measurement scale, uses it in a comparison, introduces a work budget, and then asks the learner to keep a journal against that budget. Inspection shows that the measurement and budget distinctions are available before their later uses. One return points to figure 17 when the relevant scale is in figure 26; a final diagram also contains an unexplained arrow.
The attachment repair is to correct the return and explain the relation the arrow denotes, then inspect their receiving uses. The supported measurement-before-comparison and budget-before-journal prefixes are preserved. Adding a new overview to every chapter would not follow from these two defects.
If a human reader reports a broader difficulty, retain that report under its task, prior knowledge and assistance conditions. The prepared reader’s recovery supports a different finding; a wider claim about comprehension or experienced load needs evidence about the intended audience. The confirmed return and arrow defects can still be repaired and their affected uses checked.
Suppose no target reader is available in the revision window. Return the supported local diagnosis and repair, with the reported broader difficulty still open for the wider audience.
Now change the receiving use to choosing the short route or an added overview for the next class. Suppose voluntary target-reader participation, appropriate protection and access are available before that decision: twenty reader-minutes and thirty provider/evaluator-minutes cover design, preparation, the attempt, a segment-timed burden question, feedback and interpretation. The observation can distinguish a missing connection from a prerequisite or assistance problem and thereby change which repair is used; that contribution warrants the stipulated whole burden and displaced work. Select the observation with its intended respondent, segment, timing, instrument and assistance conditions, including interference caused by asking. This completes a decision to observe, not a human-comprehension, burden, retention or learning result. The route and these availability conditions are constructed examples, not observed human Work.
HCD.19:6 - Bias-Annotation
An evaluator who knows the intended repair can unintentionally teach it through the task prompt. Preserve the uncued opportunity when that is what the specification needs. If a cue was already supplied, narrow the conclusion rather than reconstruct an imaginary first answer.
A reader’s account of source use can omit familiar prior knowledge. Inspect the material and use a discriminating condition when the attribution remains consequentially uncertain. A fluent language-model answer shows what that system produced under its reading conditions. Evidence about the intended audience is needed before generalizing to inexperienced people; identify the limit when such evidence is unavailable.
Whole-route inspection can inherit the evaluator’s preferred reading order. Judge the audience’s declared route and the sources, help and later passages it can reach where each contribution is needed. Preserve deliberate preparation and useful difficulty when the later instruction resolves the difficulty in time for its intended use.
HCD.19:7 - Conformance Checklist
Use these criteria for the requested evaluation.
- HCD19-CC1. The evaluator SHALL use a specification compatible with the material, audience, tasks, assistance, subject basis, and intended question or decision, or return the specific construction gap.
- HCD19-CC2. For a new observation, the evaluator SHALL define the task and subject criterion before seeing the response. For an existing response, the evaluator SHALL establish the applicable criterion independently of it. Where discrimination is needed, the task and criterion SHALL distinguish a consequential rival result.
- HCD19-CC3. For a new test of independent recognition, the evaluator SHALL preserve the first response before supplying answer-bearing help. For existing work, establish what evidence of that response survives and which conclusion it supports. Where assistance affects the inference, distinguish what was promised, available, received and used.
- HCD19-CC4. A whole-route conclusion SHALL use connected evidence of the complete declared route and consequential cross-part dependencies; local fragment judgements alone SHALL NOT stand in for it.
- HCD19-CC5. The result SHALL distinguish the observed contribution, source-inspected material condition, prior knowledge, support, criterion defects, and missing evidence when they change the inference.
- HCD19-CC6. A complete invocation SHALL return every selected coordinate with its value or lawful missingness, evidence, adjacent-value grounds, and limits. A bounded diagnostic SHALL declare its narrower scope.
- HCD19-CC7. A repair return SHALL identify the defect, complete affected reach, proposed change, protected contributions, and discriminating next use for E.23.
- HCD19-CC8. A repeated use SHALL preserve assistance and exposure conditions, reuse unaffected evidence, and distinguish supported continuation from stronger human performance, transfer, retention, or causal claims.
HCD.19:8 - Common Anti-Patterns and How to Avoid Them
| Misuse | Consequence in use | Better move |
|---|---|---|
| Correct the answer before saving it | The observation of first recognition is lost. | Preserve the first answer, then record the correction and help separately. |
| Call every failed task a material defect | A prerequisite gap or invalid criterion is misdiagnosed. | Compare the promise, subject criterion, and available support before selecting the repair. |
| Call every correct answer material adequacy | Expert rescue or answer-bearing help disappears from the inference. | Trace the material contribution and retain its limits. |
| Repair a central contradiction with a footnote | Dependent examples and tasks still teach or reward the wrong relation. | Identify the smallest connected repair that would correct the use. Carry it out if the task calls for repair. |
| Rebuild the whole course for a wrong return | Useful explanations and established prefixes would change without supporting grounds. | Locate the wrong link and the use that depends on it. Correct them when a repair is requested and follow any other uses dependent on the change. |
| Remove all initial difficulty | A useful preparatory attempt may be lost. | Inspect whether subsequent instruction works with and resolves the attempted alternatives. |
| Equate structural complexity with human burden | A property of the rendering is substituted for an unobserved human experience. | Retain the structural diagnosis and the person’s reported difficulty with their respective grounds. Use applicable observations of the relevant people, task and conditions for a human-burden claim. Choose a new qualified observation only if it is feasible and its attainable contribution warrants the full burden. |
HCD.19:9 - Consequences
The evaluation can establish the material’s contribution, locate a defect or identify the evidence still needed. The evaluator can distinguish a bad criterion from reader error, recognize a false explanation despite an expert’s success or repair a local reference while preserving a working route. First and helped responses support different claims.
Reader participation costs time and can change the reader’s knowledge. Reuse compatible observations and select new work for its attainable contribution to the receiving use, including reader and provider burden. Source inspection can close a narrow subject defect; missing audience evidence limits the broader claim without erasing that result. Repeated expert agreement cannot fill the audience-dependent gap.
HCD.19:10 - Architectural Rationale
Representative reader work makes a material’s contribution visible in use. It remains only one part of the inference: the subject criterion, source inspection, prior knowledge, and support arrangement determine what a response can establish. Keeping those relations separate explains why the same correct answer may accompany either a good explanation or an expert’s repair of a bad one.
Use HCD.18 to obtain a suitable specification when one is missing or needs revision. Apply a compatible specification from any qualified source. For improvement under E.23, supply the material diagnosis and affected-use test. Use NSTD to examine relations across a narrative when those relations matter.
A change in the material and a change in a person are different results. HCD.11–HCD.13 retain the stronger human evidence questions, while HCD.19 can provide a useful bounded contribution without waiting for every later outcome.
HCD.19:11 - SoTA-Echoing
The practice question is how to distinguish useful instruction and help, identify any repair, and qualify claims about learning. Compared with a single post-reading score, inspecting the material–task relation can reveal a subject defect despite a reader’s successful correction. Preserve evidence of the first attempt when first recognition matters. Following a response into a later task can test a warranted continuation, at the cost of obtaining and interpreting that further observation.
| Practice choice | Adopt, adapt, or reject; effect on this pattern | Source role, limits, and reopen condition |
|---|---|---|
| Preserve help when interpreting performance | Adapt assistance-sensitive comparison in sections 4.2 and 4.6 and the M0/M1 case. A result obtained with the answer in the worked example is still useful, but cannot establish independent acquisition. | Bastani et al. (2025) provide failure evidence for inferring later independent performance from assisted success in a particular school-mathematics AI arrangement. Its safeguards were bundled; this does not identify a universal mechanism. Reopen when the promised equipped-work arrangement or later independent claim differs. |
| Follow the relation the reader must construct across parts | Adapt principle/example integration and self-explanation in sections 4.3–4.4. Local recall is an inadequate comparator when the promised work combines separated concepts. | McCrudden et al. (2022) and 2024 contribute bounded topic-specific integration evidence. Their prompts count as support; they establish no universal principles-first sequence. Reopen for a different audience, relation, or route that makes another order preferable. |
| Preserve purposeful preparation while repairing unsupported difficulty | Adapt problem solving followed by instruction in section 4.6; reject both unconditional simplification and unqualified “productive failure.” | Sinha and Kapur (2021) synthesize conditional effects, largely in mathematics and physics. The practical requirement is that later instruction resolves the attempted alternatives. Reopen when that connection is absent or a different learning task changes the comparison. |
| Choose a representation repair by its function | Adapt mechanism-specific multimedia design in section 4.3 and the local-route case. A signal, correspondence, or pacing repair can target a difficulty without adding decoration throughout the course. | Noetel et al. (2022) supply a synthesis of multimedia design research, not proof of whole-book continuity. Reopen if the repaired representation still leaves the required relation unavailable. |
| Keep structural diagnosis separate from experienced burden | Adopt appropriate respondent and timing conditions for a selected human-burden observation; section 5.4 first retains the human report and narrows the untested claim, then selects new observation only for its obtainable, worthwhile contribution. | Krieglstein et al. (2025) provide failure evidence about retrospective load timing in studied materials. They supply no conversion from narrative epiplexity to human burden. Reopen for a changed segment, timing, instrument, or respondent population. |
HCD.19:12 - Relations
- HCD.18 can supply a suitable specification when the current one is absent or incompatible.
- HCD.6, HCD.7, and HCD.9 can supply representative practice tasks, feedback arrangements, and support for practice. An evaluator can also use equivalent qualified inputs from other sources.
- HCD.11, HCD.12, and HCD.13 govern human performance, unfamiliar-work transfer, and retention evidence. The observations obtained here remain qualified contributions to those different questions.
- NSTD.6 separates current narrative-product, source-fidelity, and construction-history evaluation. NSTD.8 supplies attachment and available-prefix reconstruction when narrative learning-route use is live.
- E.22 supplies comparison among material alternatives under a common specified basis. For improvement under E.23, supply the limiting defect, useful contributions to protect and a test of the affected use.
- Use F.19 to repair a defective technical passage once its subject meaning is understood. Correct the wording and follow any changed relation to the tasks and passages that rely on it.