{"schema_version":1,"experiment_id":"eoa_inverse_innovation_exp03_full320_20260801","cell_id":"deadweight_loss_reduction__statistics_experimental_design","trajectory_id":"R","attempt_index":0,"archetype_slug":"deadweight_loss_reduction","domain_slug":"statistics_experimental_design","decision":"CANDIDATE","problem_id":"equal_alpha_allocation_blocks_priority_weighted_inference","causal_lever_id":"prespecified_priority_weighted_alpha_allocation","proposal":{"problem":"In experiments testing heterogeneous hypotheses under a fixed familywise Type I error budget, an equal or stale allocation of alpha can give low-decision-value hypotheses the same inferential capacity as critical hypotheses. When multiplicity allocation rather than sample scarcity is binding, this can suppress power for consequential comparisons without providing additional error protection.","actors_substrate":["experimental units and participants","study investigators","statisticians and data-monitoring bodies","decision-makers using study results","groups represented by lower-priority hypotheses"],"observable_state":"Under an equal-alpha design, prespecified high-priority hypotheses have materially lower power than under valid alternative allocations, while alpha assigned to low-value or weakly plausible hypotheses contributes little expected decision value.","consequence":"Scarce observations and error probability are allocated away from comparisons with greater prespecified decision value, producing avoidable false negatives, delayed decisions, or additional sampling while the protected Type I error limit remains unchanged.","affected_objective":"Increase expected decision-relevant information per ethical sample while preserving familywise error control, prespecification, interpretability, and fair coverage.","structural_mapping":[{"archetype_element":"value-blocking wedge","domain_realization":"A uniform or inherited rule divides a fixed Type I error budget without regard to prespecified differences in decision importance or evidence.","claim_kind":"HYPOTHESIS"},{"archetype_element":"blocked beneficial activity","domain_realization":"Detection of true, consequential effects and timely evidence-based decisions is impeded by unnecessarily low thresholds for priority comparisons.","claim_kind":"HYPOTHESIS"},{"archetype_element":"idle or low-value allocation","domain_realization":"Inferential capacity is committed to comparisons whose rejection would have relatively little prespecified decision consequence.","claim_kind":"INFERENCE"},{"archetype_element":"protected constraint","domain_realization":"Strong Type I error control, outcome-independent design choices, transparent reporting, and minimum coverage for affected groups remain fixed.","claim_kind":"HYPOTHESIS"},{"archetype_element":"redesign","domain_realization":"Allocate alpha using preregistered priority weights or a gatekeeping structure derived without access to study outcomes.","claim_kind":"HYPOTHESIS"},{"archetype_element":"incidence","domain_realization":"Priority comparisons may gain power while lower-weighted comparisons lose power; participants and downstream users bear consequences of both errors and delayed evidence.","claim_kind":"INFERENCE"},{"archetype_element":"bounded reversibility","domain_realization":"Validate by simulation and archived-design replay before using the rule in one prospectively registered hypothesis family with automatic reversion.","claim_kind":"HYPOTHESIS"}],"component_map":[{"component":"Distortion Map","status":"direct","domain_realization":"Trace how the current multiplicity rule maps each hypothesis to its alpha share and resulting power."},{"component":"Protected Constraint Safeguard","status":"direct","domain_realization":"Hold the target familywise error rate, prespecification, reporting duties, and minimum protected comparisons constant."},{"component":"Surplus Estimate","status":"adapted","domain_realization":"Estimate gains in decision-weighted power, expected correct decisions, or samples avoided, with uncertainty."},{"component":"Affected-Party Incidence Map","status":"direct","domain_realization":"Show power and decision consequences for every hypothesis, participant group, and downstream user."},{"component":"Redesign Lever","status":"direct","domain_realization":"Use outcome-independent weighted alpha allocation or prespecified gatekeeping."},{"component":"Distributional Review","status":"direct","domain_realization":"Review which scientific questions and represented groups lose inferential opportunity."},{"component":"Behavioral Response Model","status":"adapted","domain_realization":"Model strategic inflation of priority labels, hypothesis splitting, and migration into nominally privileged families."},{"component":"Implementation Boundary","status":"direct","domain_realization":"Limit the first prospective use to one registered family, fixed weights, and named analyses."},{"component":"Monitoring and Rebound Check","status":"direct","domain_realization":"Audit error control, achieved power, priority-label changes, reporting completeness, and spillover into exploratory analyses."},{"component":"Rollback or Adjustment Rule","status":"direct","domain_realization":"Revert to the approved equal-allocation design if validity, fairness, or prespecification checks fail."},{"component":"Cost–Benefit Assessment Frame","status":"adapted","domain_realization":"Compare decision-weighted gains against lost power, complexity, audit burden, and gaming risk."},{"component":"Price-Wedge Diagnostic","status":"adapted","domain_realization":"Treat alpha as a constrained allocation budget only diagnostically; confirm multiplicity allocation, not low sample size or noisy measurement, binds power."},{"component":"Friction Source Breakdown","status":"adapted","domain_realization":"Decompose low power into sample size, variance, effect size, multiplicity burden, and allocation-rule contributions."},{"component":"Compensating Adjustment Plan","status":"adapted","domain_realization":"Provide alpha floors, separate exploratory status, or additional sampling for adversely affected comparisons where justified."},{"component":"Legitimacy and Authority Review","status":"direct","domain_realization":"Require approval from the study's statistical and scientific authorities and any ethics or regulatory body governing the protocol."},{"component":"Sensitivity Analysis","status":"direct","domain_realization":"Vary priority weights, effect distributions, dependence, utility assumptions, and group-level floors."},{"component":"Pilot or Sunset Path","status":"direct","domain_realization":"Use simulation and archived replay, then one sunset-bound prospective application."}],"mechanism_dispositions":[{"slug":"congestion_or_capacity_pricing_adjustment","disposition":"considered_rejected","contribution_type":"NONE","adaptation_or_rejection":"Alpha is not priced by real-time demand, and outcome-responsive repricing would threaten prespecification.","counterfactual_removal":"No change; dynamic pricing is absent from the causal chain."},{"slug":"cost_benefit_assessment_protocol","disposition":"selected_supporting","contribution_type":"TEST_DESIGN","adaptation_or_rejection":"Translate welfare into decision-weighted power while retaining nonmonetary ethical and distributional constraints.","counterfactual_removal":"The pilot could still run, but priorities and tradeoffs would lack a comparable decision rule."},{"slug":"distortion_reduction_review","disposition":"selected_load_bearing","contribution_type":"CORE_CAUSAL","adaptation_or_rejection":"Diagnose whether alpha allocation is the binding avoidable wedge and separate it from legitimate error control.","counterfactual_removal":"The proposal could mistake inadequate sampling or necessary multiplicity protection for removable distortion."},{"slug":"impact_assessment_table","disposition":"selected_supporting","contribution_type":"OPERATIONAL","adaptation_or_rejection":"Record hypothesis- and group-level power gains, losses, safeguards, and monitoring triggers.","counterfactual_removal":"Aggregate gain could conceal concentrated loss of inferential coverage."},{"slug":"matching_improvement_program","disposition":"considered_rejected","contribution_type":"NONE","adaptation_or_rejection":"No willing-party pairing failure is posited.","counterfactual_removal":"No change; matching is not causal."},{"slug":"permit_or_approval_streamlining","disposition":"incompatible","contribution_type":"NONE","adaptation_or_rejection":"Approval delay is not the proposed source of lost power.","counterfactual_removal":"No change."},{"slug":"price_control_redesign","disposition":"incompatible","contribution_type":"NONE","adaptation_or_rejection":"Multiplicity thresholds are error-control parameters, not administered market prices.","counterfactual_removal":"No change."},{"slug":"quota_or_allocation_rule_review","disposition":"selected_load_bearing","contribution_type":"CORE_CAUSAL","adaptation_or_rejection":"Adapt the quota review to a fixed alpha budget, preserving the cap while changing its prespecified allocation.","counterfactual_removal":"There would be diagnosis but no lever connecting the wedge to a valid reallocation."},{"slug":"regulatory_simplification_pilot","disposition":"selected_supporting","contribution_type":"TEST_DESIGN","adaptation_or_rejection":"Adapt to a simulation-first, single-family prospective pilot with validity monitoring and expiry.","counterfactual_removal":"Live adoption would lack a bounded evidence-producing path."},{"slug":"sunset_clause_review","disposition":"selected_supporting","contribution_type":"SAFETY_GUARDRAIL","adaptation_or_rejection":"Make the experimental allocation rule expire unless its validity and incidence justify renewal.","counterfactual_removal":"An initially bounded exception could persist without affirmative review."},{"slug":"tariff_fee_or_toll_redesign","disposition":"incompatible","contribution_type":"NONE","adaptation_or_rejection":"No monetary authority charge is involved.","counterfactual_removal":"No change."}],"causal_chain":["Equal allocation divides a fixed alpha budget without reflecting preregistered decision priorities.","Priority hypotheses receive stricter thresholds than necessary to preserve the same familywise error cap under a valid alternative allocation.","Their power and expected decision value decline while lower-value comparisons retain equal capacity.","Outcome-independent weights reallocate, rather than enlarge, the error budget.","Simulation verifies error control and estimates hypothesis- and group-level incidence.","If safeguards hold prospectively, decision-weighted power rises without relaxing the protected error rate."],"baseline":"Equal Bonferroni-style allocation across all hypotheses in the registered family, with unchanged sample size and complete reporting.","nearest_rival":"Increase sample size while retaining equal alpha allocation; it can restore power but consumes more participants and does not test whether the allocation rule is the binding wedge.","authority_safety":{"affected_parties":["participants exposed to study procedures","scientific and statistical investigators","communities represented by each hypothesis","ethics and regulatory reviewers","downstream decision-makers"],"decision_authority":"The study sponsor and accountable scientific-statistical governance body, with ethics or regulatory approval wherever the registered protocol or participant risk is affected.","authorized_first_step":"Using synthetic data and locked archived designs, compare equal allocation with several prespecified weighting rules across plausible effects and dependence structures; verify familywise error, power incidence, and gaming sensitivity before any live use.","excluded_actions":["choosing or changing weights after viewing outcomes","raising the approved familywise error rate","suppressing low-weighted results","removing minimum protections for safety or represented groups","prospective rollout before simulation validity checks"],"halt_rollback":"Halt if simulated or prospective error exceeds its tolerance, weights prove outcome-informed, a protected comparison falls below its approved floor, reporting becomes incomplete, or material group harm appears; revert to the approved equal-allocation analysis and report both the trigger and deviation."}},"negative_tests":{"strongest_counterevidence":"The case weakens if priorities cannot be specified independently of outcomes, if comparisons have genuinely symmetric value, or if dependence and effect uncertainty make weighting reduce robust power while adding gaming and interpretive complexity.","analogy_break":"Alpha is a mathematical error allowance, not transferable economic capacity or mutually owned property. Reallocation can change the meaning and robustness of inference, and no aggregate 'surplus' can override validity, ethics, or equitable scientific coverage.","failure_condition":"The redesign fails if apparent gains arise from optimistic effect or utility assumptions, invalid dependence modeling, post-outcome priority changes, or unacceptable power loss for protected comparisons.","problem_falsifier":"The diagnosed problem is absent if equal allocation is not binding—each planned comparison already meets its power requirement, priorities are substantively symmetric, or low power is instead caused by sample scarcity, measurement variance, or implausibly small effects.","intervention_falsifier":"At the same verified familywise error rate and sample size, the weighted rule fails if preregistered simulations and prospective evaluation show no robust improvement in decision-weighted power, or show unacceptable hypothesis- or group-level losses relative to equal allocation.","risks":["priority labels may encode institutional bias","investigators may game families or weights","utility estimates may falsely monetize contested scientific values","lower-weighted questions may become systematically underpowered","added complexity may reduce auditability or replicability","misspecified dependence may invalidate error control"]},"null_rationale":null,"classification":{"candidate_kind":"DOMAIN_TRANSFER","prior_art_status":"UNSEARCHED","evidence_maturity":"HYPOTHESIS"},"revision_change_log":{"revision_kind":"ORIGINAL","prior_problem_id":null,"prior_causal_lever_id":null,"problem_changed":false,"causal_lever_changed":false,"conceptual_changes":[],"operational_changes":[],"repairs_addressed":[]},"confidence":0.78,"generator_notes":"Closed-book structural candidate. Domain-specific prevalence, effect magnitude, and prior art were not established by the packet."}