Practical entries

Research Method Practice helps a practitioner turn an uncertain question into a supported answer, an informative limitation or a decision about further inquiry. The methods connect through results: a question determines what a design must distinguish; operational choices preserve that comparison; actual material supports an analysis; criticism or synthesis can change the claim returned to its user. A later contribution can also expose a failed earlier premise.

These are selected examples of that cooperation, not a catalogue or a prescribed research sequence. Start with the contribution whose inputs you have. Use the Table of Contents for another question and the Preface for the repertoire’s reasons and shared conditions. You can ask an assisting agent to explain a pattern or comment on your case in the language of your work, without framework jargon.

A short use can finish at RMP.1 — Decide Whether a Question Needs Research and When to Stop: if an applicable source already answers the receiving question, return its answer, source and relevant limits. No study or new brief is needed. If reading leaves an unresolved knowledge question, establish the non-routine choice, inspectable question-to-result connection and bounded contribution that make it research.

The examples below are constructed applications with supplied professional premises. Research can change the evidence used in a maintenance, environmental or other decision; its responsible practitioner retains the decision and its authority.

RMP-SEAL — Separate routine qualification from a mechanism-validity inquiry

• Situation: A seal engineer needs to know whether an accelerated test supports a proposed use in pump maintenance, but the relation between laboratory degradation and service conditions may be unresolved.
• Question: Which contribution can answer the evidence question, and what must change if its comparison or transfer premise fails?
• First useful result or blocker: An applicable existing answer or return to routine qualification; otherwise a bounded research question and a decision about whether an obtainable inquiry can answer it.
• Start with: RMP.1 for an unsettled question. Use RMP.2 — Choose a Criticism-Bearing Research Design or Stop directly when a compatible question, source account and useful stopping condition are already supplied.
• Stop or return: Finish the answer the evidence supports. Return a changed question to RMP.1, a failed design premise to RMP.2, and a maintenance choice to its responsible authority.

1. Separate routine qualification from a surviving knowledge question. In the Pump-P4 application, an already qualified acceleration relation, test method and acceptance criteria permit an engineering or assurance return. If current evidence already establishes that the proposed acceleration changes the degradation mechanism, return that answer. If the relation remains unresolved against serious alternatives, RMP.1 specifies what would retain, narrow or defeat it within the named material and operating envelope.
2. Use that question to choose the inquiry or its stopping point. RMP.2 distinguishes sufficient existing evidence, a justified selected design, a selection prevented by a missing result, and a completed decision to undertake no new study while the claim remains unresolved. Suppose the measurement specialist has not yet established whether the marker distinguishes the rival mechanisms. If its plausible results select different strategies, the experimental design is not yet selected. A decision to obtain that missing result remains separate. Check combined demands as well: two necessary tasks, each requiring four to six non-overlapping investigator-days, cannot be treated as fitting ten days while six plus six remains possible.
3. Preserve the selected comparison in the actual operations. Once a suitable marker, measurement procedure, material plan, safe conditions and access are supplied, RMP.3 - Operationalize the Research Design without Replacing the Question specifies specimen identity and allocation, observation timing and the comparison’s use. Three readings of one seal remain three readings, not three independent seals. If the specialist then finds that the marker cannot distinguish the mechanisms, return that failed premise to RMP.2. Substituting a convenient wear score needs its own relation to the question.
4. Give analysis the actual material and its limits. RMP.4 - Preserve a Criticism-Ready Research Trace connects specimen P4-07’s handling damage, absent observation and exclusion decision to the remaining material. RMP.5 - Analyze and Criticize the Research Claim under Its Method uses that history to examine the inference; an absent reading alone does not determine the direction of bias. If a proposed field claim instead rests on a simulation, reproducing its reported curve leaves the model-to-field relation open. RMP.6 - Choose a Credibility Test That Bears on the Research Claim, or Answer without New Testing asks whether an obtainable mechanism or field comparison can change that conclusion enough to warrant its burden. Selecting the comparison supplies no result from it.
5. Combine contributions only for the question they jointly answer. In the application’s corpus branch, RMP.7 - Synthesize a Research Evidence Corpus finds that reports A and A2 describe one test, B an independent compatible test, and C service replacement under different conditions. A2 adds information but no independent estimate. For the laboratory mean-lifetime difference, the supplied independent estimates are 14 hours with standard error 4 and 8 hours with standard error 3. Under the case’s justified common-effect model and normal approximation, inverse-variance weighting gives 10.16 hours, standard error 2.4 and an approximate 95% interval of 5.46–14.86 hours. C remains useful for its different service question.
6. Return the consequence to the claim that uses it. RMP.8 - Use Research Results to Retain or Revise an Episteme’s Claims and Return Bounded Evidence locates the claim that depends on an analysis or synthesis result. In its accompanying model example, a separate supplied analysis supports a fitted dry-laboratory wear relation; the synthesis does not establish that the same rate applies in wet service. Retain the laboratory fit, but withdraw the assertion of demonstrated transfer and the service replacement-time prediction that depends on it. Return that missing premise to the reliability receiver. The engineering wear limit and the receiver’s maintenance decision retain their own grounds.

A supplied answer can enter at analysis, credibility comparison or synthesis without recreating the preceding work. If the common-effect premise fails, revisit the synthesis interpretation; if qualified field evidence later becomes available, reconsider the transfer question rather than repeating an unchanged laboratory calculation.

RMP-CITIZEN — Investigate a stream anomaly without an affiliation gate

• Situation: Residents have downstream salinity records, a proposed explanation and material from other catchments. A comparison they intended to make becomes unavailable.
• Question: What survives the lost comparison, and how should that limit change the briefing used by an environmental decision maker?
• First useful result or blocker: A supported account of the observed pattern and the specific comparison or source relation still missing; where the research conditions hold, an appropriate inquiry or qualified no-new-study answer.
• Start with: RMP.1 if monitoring and research remain confused. With a selected field design, start at RMP.3; with inherited material, start at RMP.4.
• Stop or return: Preserve the supported observations and return the unsupported attribution. Site access, measurement qualification, community data authority and environmental response remain with their responsible participants.

1. Keep the research question distinct from threshold monitoring. In the ResidentStream-S17 application, established sampling under a supplied threshold and response rule can finish with a monitoring result. An unresolved explanation involving road-salt transport, seasonality or sensor drift needs a question that can discriminate the live alternatives. Citizen or academic affiliation does not decide that distinction.
2. Make the selected comparison recognizable in each observation. Given qualified measurement, safe access and suitable comparison periods, RMP.3 identifies the site, period, hydrological conditions, sample and procedure. The laboratory may supply its qualified measurement result. The residents connect that result to their site-period comparison.
3. Propagate the lost condition. Suppose the reference reach becomes inaccessible during the informative thaw while downstream observations remain possible. RMP.4 preserves those observations, the missed site-period and its reason. A later dry-weather reference sample does not restore the paired thaw observation. Return the lost comparison to RMP.2; keep the actual history and any justified narrower use.
4. Return the bounded claim. RMP.5 may support the downstream pattern while leaving its source unresolved. RMP.6 considers a later matched observation only when it could answer the new question and is obtainable and worthwhile. The current limited answer can finish without commissioning it.
5. Let other sources contribute without erasing the local gap. RMP.7 can compare catchment studies under their actual conditions, but another catchment’s similar chart does not recreate the missed local event. RMP.8 corrects a briefing that attributed the local anomaly beyond its evidence, preserving the downstream finding and carrying the causal uncertainty to its actual recipient.

The historical application shows the same dependence between synthesis and correction with different material. Several histories repeat one memoir; an authenticated record dates the decision to close a factory before the protest said to have caused it. That evidence defeats the proposed cause of the earlier decision. It does not settle the decision’s cause or the protest’s effect on later implementation. RMP.8 returns the identified correction to the responsible museum editor without changing it into a claim about another event.

RMP-CORRECTION — Repair an analysis and return its consequence without discarding a suitable method

• Situation: A completed analysis contains a calculation defect, and a later change in the observations may also challenge the method’s fit.
• Question: Which object needs correction now, and when does a different method comparison become necessary?
• First useful result or blocker: A verified implementation correction with the affected analysis and claim returned to their users; under changed receiving conditions, a supported method decision or the still-unresolved qualification.
• Start with: RMP.9 — Decide Whether and How to Improve Research Methods and Practice, using the actual difficulty and receiving conditions.
• Stop or return: Retain a suitable method after a determinate repair. A selected comparison under changed conditions remains a plan until its results exist.

1. Locate the defect and retain its scope. In the calculation application, the supplied method is suitable for the independent observations 2, 4, 6 and 8 and specifies sample variance with denominator n−1. Their mean is 5 and squared-deviation sum is 20. The implementation divides by n, reporting 5 instead of 20/3. RMP.4 identifies that computation and the analyses that used it.
2. Correct and verify the affected result. RMP.9 retains the method and repairs the denominator. RMP.5 recalculates the dependent analysis; RMP.8 returns any changed uncertainty or conclusion to the actual use that relied on it. The known-answer check settles this correction without new participants or an efficacy study. Unaffected analyses retain their existing support.
3. Reopen fit when the question changes. Now suppose the programme needs an equally weighted mean across independent sites, with unequal numbers of dependent observations within each site. The old pooled-observation analysis no longer answers that target under the case premises. RMP.9 compares candidates for the same site-level quantity. In the changed-situation case, a qualified statistician, existing simulation code and a two-day allocation make a useful comparison attainable: a t interval from independent site means versus resampling whole sites. The selected comparison is the result so far; the winning procedure and its qualification remain unresolved until actual results exist. Retain the restriction on the unsupported uncertainty claim and reconsider the acquisition if access, capability or worth changes.