Part I - Qualify the Question and Choose a Design or Stop

RMP.1 — Decide Whether a Question Needs Research and When to Stop

Type: Research Method Practice practitioner method pattern

RMP.1:0 - Use This When

Use this pattern when a practitioner receives a request to research an issue and needs to decide whether existing knowledge is enough or a new study is needed. Begin with the question, its intended use and the available material. Use a suitable existing source answer, or prepare one through §4.3, and stop when it is sufficient. If the request belongs to routine professional work, use §4.4 to identify the needed result and who can supply it. Produce a ResearchQuestionBrief only when all four research-membership conditions pass.

Recognition cues include a favored answer looking for support, a request for “more data,” a surprising observation, a test whose relation to field use is disputed, a literature search that may already be sufficient, or a question that survives routine recovery. These cues justify opening the boundary question; they do not establish research. Affiliation, job title, grant, venue, publication plan, tool choice, or information collection establishes none of the four conditions.

The practical gain is an answer from existing knowledge or a research question that states what remains to be learned and when enough has been learned. Use RMP.2 only after a compatible ResearchQuestionBrief and qualified current frame exist and the design disposition is still open.

Do not use RMP.1 to select a research design, prescribe an instrument, authorize access or risk, decide an engineering or policy action, or repair a reusable professional Method. Those decisions remain with their direct owners.

RMP.1:0.1 - Working Terms

Work that changes an episteme and Work that changes an operating, physical, social, or administrative state can be connected, but they are not the same Work. A research result may inform a later action; it neither performs nor authorizes that action.

RMP.1:1 - Problem Frame

Many activities appear on both sides of the research boundary. A laboratory test can execute an already qualified engineering Method or challenge the mechanism that makes the test relevant. Repeated stream measurements can implement routine monitoring or investigate an unexplained pattern and rival causes. Reading sources can retrieve an already available answer or form one part of a criticism-bearing synthesis. Interviews can support routine evaluation or a situated inquiry whose interpretations and negative cases remain open.

The boundary is therefore relational. It depends on what is unresolved relative to a qualified current frame, what episteme may change for which use, which non-routine choice and criticism are required, how the Work will remain inspectable, and what kind of result will be returned. Neither a prestigious venue nor an informal citizen setting changes those relations.

A useful entry pattern must be permissive enough to inspect an ambiguous request and strict enough not to relabel every uncertainty as research. It must also preserve a cheap exit: if current source work already answers the question, starting a new study is a defect, not diligence.

RMP.1:2 - Problem

Activity lists and institutional labels create false positives: searching, measuring, experimenting, coding, surveying, or publishing is treated as research even when it only applies a qualified Method or returns an established result. Institutional gates create false negatives: independent practitioners and community groups are excluded even when they pose a genuine unresolved question and preserve a criticism-ready trace.

A vague “not research” answer is no better. It strands useful Work without identifying the result that is actually needed, the owner who can supply it, or the condition under which a knowledge question would reopen. A vague “needs more research” answer hides the same defect in the opposite direction.

Without a useful stop, the topic expands indefinitely. Without a qualified current frame, novelty means only “new to me.” Without a named receiving use, a question can accumulate evidence without ever stating what episteme change would be useful. Without a criticism demand, a favored answer can dictate the search.

RMP.1:3 - Forces

• A broad topic is easy to announce, while an answerable question needs a named receiver, phenomenon, unit or case, contrast, constraints, and stop.
• A quick source answer saves Work, while a weak or stale source cut can create false closure.
• A familiar protocol feels systematic, while routine execution does not establish an unresolved knowledge question.
• A surprising observation can open inquiry, while surprise to one performer is not qualified novelty.
• A concrete decision motivates research, while research must not inherit the receiver’s value or authority.
• Criticism must be planned early enough to matter, while selecting the actual design belongs to RMP.2.
• Traceability supports inspection, while a checklist or repeated measurement alone does not make Work research.
• Communicability makes a result criticizable, while publication, unrestricted openness, and institutional recognition are not membership conditions.
• Holder-neutral access admits corporate, public, independent, and citizen practice, while ethics, safety, rights, confidentiality, access, and domain competence remain binding.

RMP.1:4 - Solution

Start with the receiving question and attempt a sufficient answer from existing material. Use the four membership conditions when a knowledge question remains unresolved. Return the answer in the current exchange or work product; create a ResearchQuestionBrief only when all four conditions pass.

RMP.1:4.1 - Pattern-Use Unfolding

1. Fix the receiver and use. Name who needs an answer, the exact claim, model, corpus contribution, or explanation that could change, the action or later inquiry for which it may matter, and the owner of that later decision. Keep the proposed evidence use separate from permission, release, safety, policy, maintenance, or adoption authority.
2. Find what existing material can answer. Reuse an applicable source answer when one is available. Otherwise follow §4.3 from the question and accessible texts: locate relevant passages, recover their meaning and applicability, and show what they change in the answer. Compare alternatives when their differences could change that answer or the need for a study.
3. Try the source-answer stop. If the material resolves the receiving question within its intended scope, return the answer with the sources and limits needed to use it, and stop. Use the remaining steps when a research question survives.
4. Form the candidate episteme change. State what exact claim or corpus contribution could be retained, narrowed, rejected, qualified, or added. Name the phenomenon, unit, case, population or corpus, relevant relation or contrast, constraints, and smallest useful answer. A topic, task label, desired conclusion, or data-collection plan is not yet a research question.
5. Apply the four conditions independently. Establish unresolvedness, non-routine choice under genuine uncertainty, a systematic criticism-ready trace, and a bounded communicable episteme return. All four must pass. Record why the nearest plausible non-member does or does not fit.
6. Name the criticism demand, answer condition and work stops. Identify the error or rival that a useful answer must be able to expose, without choosing a design family. State what the evidence must establish for this use, which result would defeat or narrow the proposed claim, and what remains outside the inquiry. Separately state when the present work can finish with uncertainty, refusal or a missing-input return, for example because further inquiry lacks value or authority. Such a stop preserves any unmet answer condition.
7. Return one exact disposition.
   • return the already-sufficient source/result and stop;
   • return the adjacent non-research Work and its expected result to its direct owner;
   • return a research-boundary hold naming the unavailable source or owner result and retry condition; or
   • when every condition passes, return a ResearchQuestionBrief to RMP.2.
8. Preserve later authority. The brief says why an episteme change may matter. It does not choose the receiving action, select a design, authorize Work, or promise that the answer will support the desired decision.

RMP.1:4.2 - Four-Condition Research-Membership Test

These are assurance questions for membership. Recognition cues—such as an experiment, a literature review, a laboratory, a surprising graph, an academic affiliation, or the word “research”—cannot substitute for any row.

RMP.1:4.3 - Answer from Existing Sources before Starting a Study

Use this route when the question has arrived before a suitable source answer. A supplied answer can be used directly when its question, sources and applicability fit the present use. The purpose here is to establish what remains to be investigated.

RMP.1:4.3.1 - Find and interpret the relevant material

1. Say what answer is needed. Restate the difficulty in the receiver’s terms and say how an answer could change the next action. Use the texts and search facilities available; consider time or specialist help when they limit what you can resolve. Clarify a difference in the question before spending effort on the wrong one.
2. Locate passages that can answer it. Start with the available texts, their contents and their own terminology. Search the words used in the question and likely synonyms; follow a relevant passage to its definitions, conditions, examples or cited basis when they affect the answer. Use F.1 to choose sources by what their claims or limitations contribute. If one identified source already answers the question, use it directly.
3. Repair the reading where needed. When a search misses expected material, try another expression or source. Read a found passage with the surrounding context needed to interpret it; F.0.1 helps recover a consequential expression’s local meaning. A correctly understood passage may concern a different population, setting or use. State that difference and seek the material needed for this question. Show the corrected answer so the receiver can tell what the repair achieved.
4. Return what the sources support. Explain the answer and point to the passages and editions on which it depends. Keep omissions and uncertainty that affect its use visible. A.10 supplies the claim-bound source and provenance relations. If the answer is sufficient, stop. The explanation already being used may carry the whole result; a separate retrieval log is needed only when reproducing or diagnosing the search is itself part of the work.

A missing text, unknown edition or unresolved applicability limits the conclusion that depends on it. Preserve the answerable remainder and say which source or competent contribution would resolve the remaining question. Further search is useful when its possible answers could change the receiving use; the presence of an unanswered question alone does not prescribe a new study.

RMP.1:4.3.2 - Compare alternatives when their difference matters

Use this branch when existing sources offer materially different answers, or when the receiver needs the best-known current answer. Begin with the working question and the criteria that matter for its use.

Select sources by their answer-changing roles in F.1. When the result claims a best-known current answer, apply the stricter source comparison in F.1:4.2.1 and E.8:11, including serious alternatives, relevant failure evidence and application effort. A recent publication or a history of adoption can help identify a source without establishing that preference.

For each proposed answer, recover its assumptions, conditions and support. When an approach claims an improvement over earlier work, compare the limitation it repairs with those it retains. Consult the earlier source when that comparison needs it; keep historical influence distinct from evidence that the present answer works. State a trade-off when the answers improve different characteristics. The comparison may finish by identifying a question whose answer would distinguish the alternatives.

Say what this comparison changes: the answer, its applicability, the remaining research question or the useful stop. If establishing a new conclusion from a body of research requires appraisal and synthesis, use RMP.7 with its needed question and design. F.0.2 applies when the actual work is conceptual synthesis for a framework-authoring decision. A broader G.2 pack is useful when its declared CG-Frame and repeated-refresh use is needed. These are conditional continuations; a sufficient comparison can finish here.

RMP.1:4.4 - Nearest Non-Members and Exact Returns

For adjacent Work or a missing result, carry forward the exact question, available material, expected result kind, owner, and reopen condition so useful Work can continue. A supplied answer returns to its receiver as a completed answer; it does not need a next owner task.

RMP.1:4.5 - Record the Result

A passing ResearchQuestionBrief contains these positions:

When a supplied result already answers the current use, return that answer with the source, applicability and limitation information the receiver needs to use it correctly. The current exchange or existing work product may already contain the whole result. Stop with that answer: no ResearchQuestionBrief, unused card answers, separate form or justification for omitting a study is required.

If someone will rely on the answer later or for a different use, retain the exact source/result and the scope, currentness or uncertainty that can change that receiver’s reliance. Reuse an adequate existing explanation; add only the missing limitation or reason needed for that use.

For an exact non-research return or research-boundary hold, make the following relations recoverable from the available material or a short return:

RMP.1:4.6 - Frascati Criteria Mutated for This Boundary

The OECD Frascati criteria are a valuable cross-sector comparator for R&D. RMP.1 adapts them for holder-neutral, criticism-ready research practice; it does not copy a statistical classification or institutional gate.

RMP.1:4.7 - First-Hour Qualification Card

Start with prompts 1–3, using the question and available material. If prompt 3 yields a sufficient answer, return it under §4.5 and finish the card there. Prompts 4–8 apply only while a research-membership question remains unresolved. A sentence or compact row is enough where an answer still needs to be made explicit.

1. Who needs what answer for which bounded use, and who retains the later decision?
2. What do the available sources already answer? Use §4.3 when the answer still needs to be recovered or compared; retain the sources and limitations that matter for this use.
3. Does that source answer resolve the receiving question? If yes, return it with the sources and limits needed for this use, and stop here.
4. What exact episteme could change, and what phenomenon, unit, case, corpus, relation, or contrast bounds it?
5. Why is the question unresolved relative to the frame rather than merely unfamiliar?
6. Which non-routine choice and genuine uncertainty prevent ordinary Method execution from being enough?
7. How could a critic inspect and defeat or narrow the claim?
8. What would make the answer adequate for this use? Separately, when should the present work finish with uncertainty or refusal, and what would reopen it?

If an answer still needed to decide the current branch depends on an unavailable source or owner result, name that hold immediately. If a nearest non-member fits, return it immediately. An unused later prompt is neither a missing input nor a reason to reopen a completed answer.

RMP.1:4.8 - What Changes in Practice

The practitioner stops asking, “Is this activity called research?” and asks, “What exact episteme remains unresolved for this use, why does it need a non-routine criticism-bearing route, and what is the cheapest honest stop?” The output either closes the request from current evidence, moves ordinary Work to its real owner, exposes a precise supply gap, or gives RMP.2 a question it can actually design for.

RMP.1:5 - Archetypal Grounding

All cases below are constructed pedagogical cases. Their stipulated facts demonstrate the boundary method; they are not empirical claims about an actual pump, stream, archive, instrument, or authority.

RMP.1:5.1 - Corporate Seal-Test Case

A reliability engineer is asked to “research whether an accelerated seal test justifies changing the maintenance interval” for Pump-P4. The maintenance authority, safety owner, and domain engineering owner remain separate from the inquiry.

First apply the cheap branch. If the current engineering frame already supplies a qualified field failure mechanism, an applicable acceleration relation, a qualified test Method and instrument, acceptance criteria, and the exact evidence-to-use rule, then executing that test is routine qualification or validation. Return its test result to the engineering or assurance owner and the interval choice to the maintenance authority. The laboratory setting and possible publication do not make the Work research.

Suppose instead that the qualified current frame exposes this unresolved issue: the proposed high-temperature condition may change the dominant degradation mechanism, while installation damage and fluid exposure remain serious field rivals. A passing ResearchQuestionBrief can then read:

This case enters RMP.2 only with that brief and frame. If current retrieval already establishes that the acceleration relation is invalid for the named use, return that exact result and stop; a new study is not required merely to make the answer look experimental.

RMP.1:5.2 - Citizen Stream-Salinity Transfer

A resident team without academic affiliation observes an urban-stream salinity increase. If it is applying an established sampling Method, threshold, and response rule to report current state, return the observation to the domain monitoring owner. Repeated sampling and a public dashboard do not by themselves establish research.

Suppose a qualified current frame instead leaves sensor drift, road-salt transport, groundwater contribution, a changed outfall, and seasonal hydrology as serious rivals, and the community needs a bounded explanation for a named reach and period. The question may pass all four conditions: the explanation is unresolved for the use; the source, site, and comparison choices are non-routine; samples, instruments, conditions, exclusions, and interpretations can be traced and criticized; and the intended result is a communicable claim with limits rather than only an alert.

The brief names the reach, period, observations, relevant contrasts, source/currentness limits, rival accounts, criticism demand, safe access and sampling constraints, useful stop, and the public or operational receiver. Site permission, laboratory competence, environmental safety, privacy, and official response authority remain with their owners. Citizen status neither weakens nor strengthens membership. If the sensor is found invalid before the question can be qualified, return measurement repair or monitoring Work; do not hide the missing measure inside a research label.

RMP.1:5.3 - Historical Source-Critical Contrast

A local archive group asks whether a change in maintenance practice followed one directive or emerged earlier across workshop records. The question can be research without an experiment and without academic affiliation when the current frame is incomplete, rival chronologies matter, source provenance and silences need non-routine criticism, the trace is inspectable, and the result is a bounded historical claim. If one current authenticated record already answers the narrow date question, return that source result instead. Access, rights, conservation, and institutional interpretation authority remain external.

RMP.1:5.4 - Retrieval-Sufficient Stop

A product team asks to “research whether the current standard still permits configuration X.” A current question-relative source cut, needed source-local meanings, and exact evidence-use path establish the applicable edition and clause for that use, including its jurisdiction and effective-date limits. RMP.1 returns the answer and those source limits in the current exchange; the card ends at prompt 3. No membership account, interviews, survey or test is needed to complete that question.

A later team receives the answer for a proposed use in another jurisdiction. The retained edition, clause and jurisdiction limit show what the answer covers and make the new applicability question explicit. Obtain the missing applicability result from its competent source; the earlier answer does not become universal and the changed use does not by itself commission research. If the team instead disputes the clause’s empirical adequacy for a materially new case, that new question needs its own current frame and four-condition test.

RMP.1:5.5 - Recover an Answer from the Texts

A product team asks to research whether instruction M still permits configuration X on housing V2. It has a 2021 instruction, the 2025 revision and a 2025 correction notice. All are stipulated source texts for this constructed case.

Searching the old instruction for “permitted configurations” returns X for V1. The new contents instead use “installation combinations”; its table lists X with V2. The adjacent note limits that row to dry service. The correction notice changes a different combination. Reading the table and its note changes the answer from “only V1” to “the 2025 instruction includes X with V2 for dry service.” The answer identifies the edition, table and note. No experiment is needed to establish what that instruction says.

The team then reveals that it means wet service. The earlier answer remains useful within its dry-service scope; it does not resolve this new applicability question. The next useful result is the applicable instruction or an engineering judgement about wet service. If the team instead asks whether the instruction’s claim holds in a materially new operating setting, qualify that empirical question under the four conditions before choosing a study.

RMP.1:5.6 - Compare the Answers that Existing Reports Can Supply

Return to the seal-test question in §5.1. One stipulated report supports a temperature-acceleration relation for a named material in water. A later report concerns another fluid and reports a different dominant degradation mechanism. Both are relevant to the receiving question; their different conditions prevent a direct ranking by publication date or test duration.

The comparison retains the first report’s supported relation in water and the second report’s warning about mechanism change in the other fluid. It changes “Which report gives the best acceleration factor?” into “Does the proposed test preserve the field mechanism for our material and fluid?” If an applicable existing study answers that question, return its answer and stop. Otherwise the remaining mechanism question and the criticism it requires become inputs to the research-membership test. A new study is chosen only through the subsequent design decision.

RMP.1:6 - Bias-Annotation

RMP.1:7 - Conformance Checklist

Apply these questions to the branch actually taken. The episteme-change and membership questions govern a ResearchQuestionBrief; a supplied answer, adjacent-Work return or research-boundary hold does not complete unused brief positions.

• RMP1-01: A recognizable ambiguous request, first move, practical gain, stop, and non-use boundary are visible before the detailed method.
• RMP1-02: The named receiver, bounded use, and separate later-decision authority are explicit.
• RMP1-03: The current frame is question-relative and records exact supplied results, source snapshot, omissions, and currentness limits.
• RMP1-04: Serious rivals or lineage are present when they can change the question; a source count is not used as coverage.
• RMP1-05: A sufficient supplied answer completes both the method and the card before unused brief positions; it retains the source and limits needed for the receiving use without a separate no-study form.
• RMP1-06: The intended episteme change is exact enough to be retained, narrowed, rejected, qualified, or added.
• RMP1-07: The bounded question names the phenomenon, unit, case, population or corpus, relevant relation or contrast, constraints, and useful answer.
• RMP1-08: Unresolvedness is established relative to the qualified frame and named use, not to the performer’s memory.
• RMP1-09: A non-routine design, interpretation, synthesis, or credibility choice and genuine route/answer uncertainty are explicit.
• RMP1-10: The intended trace can expose sources, premises, performed moves, evidence, deviations, limits, and stops to criticism.
• RMP1-11: The intended result is a bounded communicable episteme contribution, not only action, compliance, acceptance, recognition, or private belief.
• RMP1-12: All four membership conditions pass independently before a ResearchQuestionBrief is returned.
• RMP1-13: The nearest plausible non-member is named and its difference is stated.
• RMP1-14: Routine qualification, monitoring, troubleshooting, audit/evaluation, source use, and ordinary Method application have exact owner/result returns and reopen conditions.
• RMP1-15: Missing or stale source/results produce a named hold with blocked question, owner/interface, and retry condition.
• RMP1-16: From the available question and source texts, the practitioner can follow §4.3 to a supported answer, a useful comparison or a named remaining gap. Source search, interpretation and applicability failures lead to the corresponding repair.
• RMP1-17: The application preserves the meanings and conditions of its FPF contributions. A comparison claimed as SoTA meets F.1:4.2.1 and E.8:11; ordinary source use follows its narrower question.
• RMP1-18: The criticism demand names an error or rival to expose without selecting a design family.
• RMP1-19: The brief distinguishes the evidence needed for an adequate answer from conditions for finishing with refusal or uncertainty; any unmet answer condition and reopen condition remain visible.
• RMP1-20: Domain truth, instruments, safety, ethics, rights, access, assurance, and authority remain with their direct owners.
• RMP1-21: RMP.2 receives only a compatible brief and frame; no design, protocol, authorization, or receiving decision is implied.
• RMP1-22: Corporate and citizen or independent cases receive the same membership test, with holder-specific constraints preserved.

RMP.1:8 - Common Anti-Patterns and How to Avoid Them

RMP.1:9 - Consequences

The receiver gains either a question fit for design comparison or a cheaper, more honest result. Unnecessary studies stop early. Routine engineering, Operations, assurance, audit, evaluation, source work, and Method application keep their own results and authority. Independent and citizen practitioners are admitted by the quality of the question-to-return relation rather than institutional status.

The cost is disciplined source qualification before attractive study activity begins. Some requests remain held because a current frame, rival map, measure, domain criterion, access result, or authority result is unavailable. Some highly visible “research” work returns as ordinary practice. These are useful losses: they make the missing value and next action explicit.

RMP.1 also constrains downstream overreach. RMP.2 receives a criticism demand but no preselected design, and the eventual receiver receives evidence relevance without surrendered decision authority.

RMP.1:10 - Rationale

The four conditions separate the reason to inquire from the activity used. Unresolvedness prevents personal ignorance from masquerading as novelty. Non-routine choice distinguishes epistemic design from ordinary Method execution. Criticism-ready systematicity makes defeat and narrowing possible. A bounded communicable result distinguishes an episteme contribution from an immediate action, token, or private update.

The supplied-answer stop has equal standing with the passing branch. Research is not improved by repeating a question that qualified current evidence already answers. Exact non-member returns avoid the opposite error: adjacent practices remain valuable and can reopen RMP.1 when they expose a new mechanism, validity, transfer, measurement, source, or evidence question.

Beginning with the receiver but withholding its authority preserves relevance without advocacy capture. Stating the criticism demand before design comparison prevents a favored apparatus or prestige hierarchy from choosing the question.

RMP.1:11 - SoTA-Echoing

The working question is: How can a practitioner distinguish a genuine criticism-bearing research question from adjacent professional Work without making institutional status or one universal design the gate?

The selected current line adapts the five criteria in the OECD’s Frascati Manual 2015—novel, creative, uncertain, systematic, and transferable and/or reproducible—while respecting that the Manual is principally a methodology for R&D statistics. RMP.1 mutates “novel” into unresolved relative to a qualified frame and use, “creative” into a non-routine epistemic choice, “uncertain” into uncertainty about the answer or defensible route, “systematic” into a criticism-ready trace, and “transferable/reproducible” into a bounded communicable contribution that does not force every inquiry into replication or universal generalization.

UNESCO’s Recommendation on Open Science supports participation by citizen and community actors and recognizes research contributions beyond a narrow institutional holder. RMP.1 uses that line to remove affiliation as a membership condition. It does not turn unrestricted openness into a requirement: ethics, safety, rights, confidentiality, security, access, and community authority continue to govern what may be done or shared.

A serious alternative is an institutional or activity-list gate: classify Work by employer, funding, venue, laboratory setting, publication intent, or use of familiar “research methods.” That gate is cheap to administer but creates both false positives and false negatives and cannot tell a routine test from a mechanism challenge. RMP.1 retains those features only as discovery cues and decides membership from the question, route uncertainty, criticism-ready trace, and intended episteme return.

These sources do not decide claim truth, supply domain instruments, authorize Work, or identify one universal design. Reopen this pattern’s boundary claims when a stronger cross-sector research-classification source changes a criterion, when evidence shows that one condition misroutes a design family or holder, or when a source/currentness rule changes the qualified-frame result. Reopen only the affected claim or return, not every prior use.

RMP.1:12 - Relations

• F.1 — Question-Relative Source Selection: supplies a question-relative SourceCutNote; it does not decide research membership or relate every source meaning.
• F.0.1 — Source-Local Meaning Recovery: supplies only the source-local meanings needed by the question.
• A.10 — Evidence Graph Referring: supplies exact evidence-use and provenance paths and limits; it does not create truth, assurance, authority, or Work.
• F.0.2 and G.2: supply conceptual synthesis for a framework-authoring decision and a broad refreshable SoTA pack respectively, when those uses are current. Ordinary source-answer preparation follows §4.3.
• E.8:11: defines SoTA and its comparison roles. Section 4.3.2 applies those distinctions to what the existing research can answer; its practical consequence is a changed answer, applicability, research question or stop.
• Current Method Engineering: owns reusable Method identity, repertoire, requirements, qualification, trial, fit, worth, variants/provenance, and introduction or revision. RMP.1 owns only the research-specific boundary and question semantics.
• RMP.2 — Choose a Criticism-Bearing Research Design or Stop: consumes a compatible ResearchQuestionBrief and qualified current frame. RMP.1 does not select its design disposition.
• Domain engineering, Operations, assurance, audit, evaluation, safety, ethics, rights, access, instrument, governance, and receiving authorities: supply their own results and retain their own decisions.
• OECD Frascati Manual 2015: supplies the five-criterion comparator, adapted here rather than treated as a statutory or institutional gate.
• UNESCO Recommendation on Open Science: supports holder-neutral citizen and community participation while leaving legitimate restrictions and authority intact.

RMP.1:End

RMP.2 — Choose a Criticism-Bearing Research Design or Stop

Type: Research Method Practice practitioner method pattern

RMP.2:0 - Use This When

Use this pattern when a compatible ResearchQuestionBrief and qualified current frame already exist, but it is still unclear whether additional evidence-producing Work is useful and, if so, which design family can expose the error that matters. Begin by testing the current evidence against the brief’s answer condition: what it must establish for this use. If new Work could still be useful, compare the real design-or-stop options on question fit, feasible criticism, validity threats, constraints, expected information gain, and supplied owner results.

Return one ResearchDesignDisposition for the present selection boundary: EvidenceSufficientStop when evidence already answers the use; ResearchDesignDecision when a useful feasible strategy is selected, with only genuinely invariant PostSelectionNeed items remaining; ResearchDesignHold when a named missing result prevents a particular selection still needed; or QualifiedNoNewStudy when the current allocation question is answered by selecting no new study while preserving an insufficient or bounded answer. A design is complete only when no missing construct relation, source admissibility, measurement, access, protection, authority or other load-bearing value can overturn its family, criticism or feasibility.

The practical gain is a useful answer about the current research commitment: use sufficient evidence, select a justified study, identify what still prevents selection, or finish with a bounded answer and no new study. A selected design is a justified strategy, not a protocol, performed Work, authorization, or promise that the desired answer will obtain.

Do not use RMP.2 to qualify an ambiguous “research” request, repair a weak current frame, write operational definitions, issue safety or ethics approval, procure access, engineer a reusable Method, execute a study, analyze records, or choose the receiving action. Return those questions to their direct owners.

RMP.2:0.1 - Working Terms

Recognition is cheap: a team has a qualified question and an open choice about further inquiry. Assurance is stricter: real options address the same brief, a selected strategy has its load-bearing owner results, and the returned disposition keeps evidence adequacy separate from the current research commitment.

RMP.2:1 - Problem Frame

Research questions do not determine one universal design. The same practical use can be served by a controlled comparison, a field observation, a qualitative account of situated mechanism, historical source criticism, a computational challenge, a synthesis of an existing corpus, a mixed composition, or no new study at all. Each exposes different errors and requires different observations, sources, access, and domain judgements.

A design family can also change when one external fact changes. If an accelerated condition alters the failure mechanism, an experimental comparison may need a different relation or may cease to bear on the field claim. If a salinity sensor is invalid or a site is inaccessible, a proposed field design may collapse or give way to a source synthesis. If archival provenance is unresolved, the relevant historical contrast may be impossible. Calling such values “details to settle later” creates a fictitious design.

The opposite error is endless pre-design elaboration. Exact file names, trace serialization, local schedule placement, or other bounded implementation choices can remain after selection when they cannot change the family or criticism logic. RMP.2 needs a test that separates those needs from load-bearing unknowns.

RMP.2:2 - Problem

Method prestige often chooses the design before the question’s criticism demand is understood. Randomization, statistical significance, interviews, simulation, triangulation, systematic review, or AI-assisted synthesis becomes a ritual answer rather than one candidate relation. The resulting design may be detailed but unable to expose the error that matters.

Premature selection hides missing owner results: a construct is not related to an observable, a source is inadmissible, an instrument cannot discriminate the change, a site cannot be entered safely, participants cannot be recruited legitimately, or the authority needed for the proposed Work does not exist. The team continues into protocol work and discovers too late that a different design—or no study—was required.

Indiscriminate holding wastes the other way. If a remaining need only assigns an already-required record locator or selects a schedule within a qualified fixed window, holding the family decision adds no epistemic protection. The practitioner needs a completed branch whose boundary can be tested, not a confidence adjective.

RMP.2:3 - Forces

• Existing evidence can make a new study wasteful, while residual uncertainty alone does not show that evidence is sufficient.
• Unlike designs protect different claims, while a common score can erase the exact criticism each enables.
• Real alternatives expose favorite-method capture, while a fixed option quota can manufacture decorative comparisons.
• Domain, ethics, safety, access, and authority results are external, while their values may determine which design is possible.
• Early design selection gives direction, while a load-bearing unknown can make that direction fictional.
• Operational detail can wait, while “we will solve it later” often hides a selection-changing premise.
• Expected information gain supports proportionate Work, while uncertain gain cannot always be reduced to one probability or utility number.
• Mixed designs can connect unlike evidence relations, while adding methods does not guarantee triangulation or agreement.
• Reporting guidance improves transparency, while a checklist cannot choose or validate a design.
• The receiver needs relevance, while research must not inherit the receiver’s decision authority.

RMP.2:4 - Solution