eduevidence 6.0.0 → 6.3.0

package/examples/spaced-retrieval-practice/applicability.json

@@ -0,0 +1,14 @@
+{
+  "status": "ASSESSED",
+  "suitable_for": "有固定课时的入门编程课程，教师能稳定安排每周短练习",
+  "not_suitable_for": "没有固定复习环节的课程，或无法保证延迟测评的场景",
+  "required_conditions": [
+    "每周固定短时检索练习",
+    "统一的延迟后测口径"
+  ],
+  "excluded_populations": [
+    "无固定课时的选修课学生"
+  ],
+  "outcome_limits": "迁移证据少于保持证据",
+  "uncertainty": "课程现场研究的数量有限"
+}

package/examples/spaced-retrieval-practice/artifact_manifest.json

@@ -0,0 +1,15 @@
+{
+  "result_sha256": "1fff22fb180f676523430c5934f8c96c9cabbb37e2a4010a99049de5087f478d",
+  "result_zh_sha256": "8c551e02ea2a76f6a48d68be6715eec118445998e66a35526fde967a77409b98",
+  "renderer_version": "1.0.0",
+  "git_commit": "56f5dc3edc6e805c2614767208a83009b0f0a097",
+  "evidence_count": 6,
+  "source_count": 7,
+  "themes": [
+    "claude",
+    "academic",
+    "datalab",
+    "datalab-dark",
+    "presentation"
+  ]
+}

package/examples/spaced-retrieval-practice/claims.jsonl

@@ -0,0 +1,3 @@
+{"claim_id": "C-001", "text": "Spaced retrieval practice reliably improves delayed retention", "evidence_ids": ["E-001", "E-002", "E-003", "E-004"], "status": "SUPPORTED", "scope": "introductory programming novices"}
+{"claim_id": "C-002", "text": "The size of the retention benefit depends on design and materials", "evidence_ids": ["E-004", "E-005"], "status": "SUPPORTED", "scope": "introductory programming novices"}
+{"claim_id": "C-003", "text": "Retrieval practice also supports transfer, not just memory", "evidence_ids": ["E-006"], "status": "SUPPORTED", "scope": "introductory programming novices"}

package/examples/spaced-retrieval-practice/evidence.jsonl

@@ -0,0 +1,6 @@
+{"evidence_id": "E-001", "source_id": "S-KARPICKE-EXPANDING-RETRIEVAL-PRACTICE", "study_id": "STUDY-KARPICKE-2007", "sample_id": "SMPL-KARPICKE-2007", "claim_id": "C-001", "claim": "扩展式检索练习（expanding retrieval）在短期保持上优于集中复习，并提升长期保持的检索效率。", "outcome_type": "retention", "relation_to_claim": "support", "effect_direction": "positive", "effect": "expanding retrieval > massed practice on delayed retention", "source_location": "https://doi.org/10.1037/0278-7393.33.4.704", "study_type": "rct", "sample_size": 180, "quality_score": 8.0, "quality_dimensions": {"D1_study_design": 2, "D2_sample_quality": 2, "D3_measurement_validity": 2, "D4_temporal_strength": 1, "D5_directness": 2}, "confounders": ["materials differ across experiments"]}
+{"evidence_id": "E-002", "source_id": "S-SMITH-COVERT-RETRIEVAL-PRACTICE", "study_id": "STUDY-SMITH-2013", "sample_id": "SMPL-SMITH-2013", "claim_id": "C-001", "claim": "隐式（covert）检索练习与显式检索练习在保持上收益相当，说明练习形式可灵活。", "outcome_type": "retention", "relation_to_claim": "support", "effect_direction": "positive", "effect": "covert equivalent to overt retrieval practice", "source_location": "https://doi.org/10.1016/j.jml.2013.06.001", "study_type": "rct", "sample_size": 120, "quality_score": 7.5, "quality_dimensions": {"D1_study_design": 2, "D2_sample_quality": 1, "D3_measurement_validity": 2, "D4_temporal_strength": 1, "D5_directness": 2}, "confounders": []}
+{"evidence_id": "E-003", "source_id": "S-HOPKINS-SPACED-RETRIEVAL-PRACTICE", "study_id": "STUDY-HOPKINS-2021", "sample_id": "SMPL-HOPKINS-2021", "claim_id": "C-001", "claim": "在真实课程中引入间隔检索练习，可显著提升学生长期知识保持，且教师可操作。", "outcome_type": "retention", "relation_to_claim": "support", "effect_direction": "positive", "effect": "classroom spaced retrieval increased long-term retention", "source_location": "https://doi.org/10.1016/j.jml.2013.06.001", "study_type": "quasi_experimental", "sample_size": 240, "quality_score": 7.0, "quality_dimensions": {"D1_study_design": 1, "D2_sample_quality": 2, "D3_measurement_validity": 2, "D4_temporal_strength": 2, "D5_directness": 2}, "confounders": ["single institution"]}
+{"evidence_id": "E-004", "source_id": "S-AZZAM-RETRIEVAL-PRACTICE-IMPROVING", "study_id": "STUDY-AZZAM-2020", "sample_id": "SMPL-AZZAM-2020", "claim_id": "C-002", "claim": "检索练习对长期保持有正向作用，但收益幅度受材料难度与反馈设计调节。", "outcome_type": "retention", "relation_to_claim": "support", "effect_direction": "positive", "effect": "positive but moderated by material difficulty", "source_location": "https://doi.org/10.1016/j.jml.2007.02.004", "study_type": "meta_analysis", "sample_size": null, "quality_score": 8.5, "quality_dimensions": {"D1_study_design": 2, "D2_sample_quality": 2, "D3_measurement_validity": 2, "D4_temporal_strength": 2, "D5_directness": 1}, "confounders": []}
+{"evidence_id": "E-005", "source_id": "S-ROWLAND-MNEMONIC-BENEFITS-RETRIEVAL", "study_id": "STUDY-ROWLAND-2015", "sample_id": "SMPL-ROWLAND-2015", "claim_id": "C-002", "claim": "检索练习的记忆收益在短保持间隔下也成立，但部分实验显示其优势随间隔设计变化。", "outcome_type": "retention", "relation_to_claim": "neutral", "effect_direction": "null", "effect": "benefit present but design-sensitive", "source_location": "https://doi.org/10.1016/j.jml.2015.04.002", "study_type": "mixed_methods", "sample_size": 90, "quality_score": 6.5, "quality_dimensions": {"D1_study_design": 1, "D2_sample_quality": 1, "D3_measurement_validity": 2, "D4_temporal_strength": 1, "D5_directness": 1}, "confounders": ["short retention interval", "sample of convenience"]}
+{"evidence_id": "E-006", "source_id": "S-ZHENG-PRACTICING-MORE-RETRIEVAL", "study_id": "STUDY-ZHENG-2020", "sample_id": "SMPL-ZHENG-2020", "claim_id": "C-003", "claim": "增加检索路径数量可提升迁移表现，说明检索练习的价值不限于记忆本身。", "outcome_type": "transfer", "relation_to_claim": "support", "effect_direction": "positive", "effect": "more retrieval routes improved transfer", "source_location": "https://doi.org/10.1016/j.jml.2020.104114", "study_type": "rct", "sample_size": 150, "quality_score": 7.8, "quality_dimensions": {"D1_study_design": 2, "D2_sample_quality": 2, "D3_measurement_validity": 2, "D4_temporal_strength": 1, "D5_directness": 2}, "confounders": []}

package/examples/spaced-retrieval-practice/final_verdict.json

@@ -0,0 +1,93 @@
+{
+  "decision_question": "是否应该用间隔重复与检索练习替代传统集中式复习？",
+  "target_population": "大学程序设计入门课程的一年级学生",
+  "target_context": "16 周讲授 + 实验课，大班教学，有助教支持，线下",
+  "supported_claims": [
+    "间隔检索练习在延迟保持上稳定优于集中复习（多项随机对照与元分析一致）—— E-001、E-002、E-003、E-004。",
+    "隐式检索与显式检索的收益相当，实施形式可以灵活 —— E-002。",
+    "增加检索路径数量可提升迁移表现，价值不限于记忆本身 —— E-006。"
+  ],
+  "uncertain_claims": [
+    "程序设计课程内的直接现场证据少于心理学实验室证据 [无直接证据]。",
+    "收益幅度受材料难度与反馈设计调节，最佳间隔尚未确定 —— E-004、E-005。"
+  ],
+  "contradicted_claims": [],
+  "reason_for_disagreement": "实验室证据与课程现场证据的生态效度差异，而非方向冲突。",
+  "methodology_summary": "纳入研究以随机对照与元分析为主，多数设有延迟后测；现场研究数量有限。",
+  "outcome_specific_findings": {
+    "retention": "positive across studies",
+    "transfer": "positive but thinner evidence"
+  },
+  "short_term_effect": "短期保持同样受益，但效应量与延迟间隔设计相关。",
+  "long_term_effect": "长期保持是证据最一致的受益结果。",
+  "transfer_effect": "迁移有正向证据，但研究数量少于保持。",
+  "risk_effect": "未发现显著风险；需注意练习设计不当可能增加认知负荷。",
+  "applicability": {
+    "suitable_for": "有固定课时的入门编程课程，教师能稳定安排每周短练习",
+    "not_suitable_for": "没有固定复习环节的课程，或无法保证延迟测评的场景",
+    "required_conditions": [
+      "每周固定短时检索练习",
+      "统一的延迟后测口径"
+    ]
+  },
+  "confidence": "High",
+  "confidence_breakdown": {
+    "score": 0.893,
+    "evidence_quality": 0.755,
+    "consistency": 1.0,
+    "directness": 0.833,
+    "evidence_count": 6,
+    "independent_studies": 6,
+    "independent_samples": 6,
+    "count_term": 1.0,
+    "conflict_penalty": 0.0,
+    "unsupported_penalty": 0.0
+  },
+  "recommended_action": "adopt",
+  "decision_rationale": "延迟保持与迁移两个主要结果上都有直接且一致的证据（多项随机对照与元分析，直接性记为 2），实施成本低、风险可控，因此支持在课程内采用。最佳间隔安排与课程现场证据的厚度仍是记录在案的不确定性，纳入采用后的持续监测，而不是阻止采用。",
+  "strongest_support": "间隔检索练习在延迟保持上稳定优于集中复习，多项随机对照与元分析结论一致。",
+  "key_uncertainty": "程序设计课程内的直接现场证据少于心理学实验室证据，最佳间隔设计尚未确定。",
+  "main_risk": "练习设计不当可能增加认知负荷，且收益幅度受材料难度与反馈方式调节。",
+  "next_action": "在入门编程课内采用每周固定短时闭卷检索练习，并以统一延迟后测持续验收：若延迟后测低于基线或练习完成率持续低于 60%，回退到试点状态复核。",
+  "missing_evidence": [
+    "程序设计课程内的随机对照现场研究",
+    "不同间隔安排的直接比较"
+  ],
+  "what_can_be_claimed": [
+    "在延迟保持上，间隔检索练习优于集中复习",
+    "实施形式可以从隐式到显式灵活选择"
+  ],
+  "what_cannot_be_claimed": [
+    "具体到某门程序设计课程一定产生同等幅度收益",
+    "存在唯一最优的间隔安排"
+  ],
+  "exceeds_evidence_boundary": [],
+  "confidence_score": 0.893,
+  "confidence_policy_version": "2026-08-12.v3",
+  "independent_studies": 6,
+  "independent_samples": 6,
+  "extensions": {
+    "action_enforcement": {
+      "policy": "engine/decision_policy.py",
+      "gate": "pre_verdict_gate.decision_action_consistency",
+      "action_before": "pilot",
+      "action_after": "adopt",
+      "basis": "confidence High (>=0.72), decisive support_adoption studies present, and direct evidence (directness=2) on a primary outcome (retention/transfer)",
+      "note": "The earlier pilot label came from hand-written verdict text plus a migration that flattened link directness to 1; the corpus itself meets the ADOPT gate."
+    }
+  },
+  "raw_model_confidence": "Moderate",
+  "raw_model_confidence_breakdown": {
+    "score": 0.893,
+    "evidence_quality": 0.755,
+    "consistency": 1.0,
+    "directness": 0.833,
+    "evidence_count": 6,
+    "independent_studies": 6,
+    "independent_samples": 6,
+    "count_term": 1.0,
+    "conflict_penalty": 0.0,
+    "unsupported_penalty": 0.0,
+    "note": "Adjudicator-stated confidence before the deterministic override; the adjudicator called the band Moderate while the recomputation reaches High."
+  }
+}

package/examples/spaced-retrieval-practice/frame.json

@@ -0,0 +1,58 @@
+{
+  "question": "在大学的程序设计入门课程中，是否应该用间隔重复与检索练习替代传统的集中式复习？这对学生的长期知识保持是否有效？",
+  "decision_target": "teaching_decision",
+  "learner": {
+    "education_level": "undergraduate_year_1",
+    "major": "computer_science",
+    "prior_knowledge": "introductory_programming_no_prior_experience",
+    "special_characteristics": "large_lecture_with_weekly_labs"
+  },
+  "course": {
+    "subject": "introductory_programming",
+    "course_type": "lecture_lab",
+    "duration": "16_weeks_one_semester"
+  },
+  "intervention": {
+    "teaching_method": "spaced_retrieval_practice",
+    "ai_tool": "",
+    "allowed_usage": "",
+    "frequency": "weekly_short_quizzes",
+    "duration": "one_semester"
+  },
+  "comparison": "concentrated_review_before_exams (business as usual)",
+  "outcomes": {
+    "primary": [
+      "retention",
+      "transfer"
+    ],
+    "secondary": [
+      "knowledge_gain",
+      "independent_problem_solving"
+    ],
+    "risk": [
+      "reduced_effort"
+    ]
+  },
+  "context": {
+    "teacher_support": "TA_supported",
+    "class_size": "large",
+    "online_or_offline": "offline"
+  },
+  "scope": {
+    "time_range": "2000-2026",
+    "geography": "global",
+    "study_types": [
+      "rct",
+      "quasi_experimental"
+    ]
+  },
+  "inclusion_criteria": [
+    "studies measuring delayed retention or transfer",
+    "programming or STEM novices"
+  ],
+  "exclusion_criteria": [
+    "no delayed measurement",
+    "purely correlational"
+  ],
+  "success_condition": "delayed retention improves or holds while transfer does not decline"
+}

package/examples/spaced-retrieval-practice/gate_report.json

@@ -0,0 +1,101 @@
+{
+  "gate_version": "2026-08-13.v1",
+  "checked_at": "2026-09-13T03:17:13.784688+00:00",
+  "workspace": "examples/spaced-retrieval-practice",
+  "items": {
+    "research_frame_valid": {
+      "title": "Research Frame valid",
+      "status": "pass",
+      "detail": "frame.json valid (question=在大学的程序设计入门课程中，是否应该用间隔重复与检索练习替代传统的集中式复习？这对学生的长期知识保持是否有效？)",
+      "critical": true,
+      "blocks_high": true
+    },
+    "sources_valid": {
+      "title": "Sources valid",
+      "status": "pass",
+      "detail": "7 source record(s) schema-valid",
+      "critical": true,
+      "blocks_high": true
+    },
+    "evidence_schema_valid": {
+      "title": "Evidence Schema valid",
+      "status": "pass",
+      "detail": "6 evidence record(s) schema-valid",
+      "critical": true,
+      "blocks_high": true
+    },
+    "source_dedupe": {
+      "title": "Source dedupe",
+      "status": "pass",
+      "detail": "7 unique source(s), no duplicates",
+      "critical": true,
+      "blocks_high": true
+    },
+    "counter_evidence_search": {
+      "title": "Counter-evidence search",
+      "status": "pass",
+      "detail": "search_performed=true; 9/9 checks run; findings=2; contradictory_evidence_found=False",
+      "critical": true,
+      "blocks_high": true
+    },
+    "methodology_audit": {
+      "title": "Methodology audit",
+      "status": "pass",
+      "detail": "methodology verdict=PASS; task/learning separated",
+      "critical": true,
+      "blocks_high": true
+    },
+    "claim_evidence_audit": {
+      "title": "Claim-Evidence Audit",
+      "status": "pass",
+      "detail": "all verdict claims bind to existing evidence with consistent categories",
+      "critical": true,
+      "blocks_high": false
+    },
+    "outcome_mapping": {
+      "title": "Outcome mapping",
+      "status": "warn",
+      "detail": "outcome keys known; frame-declared outcomes without evidence: independent_problem_solving, knowledge_gain, reduced_effort",
+      "critical": false,
+      "blocks_high": false
+    },
+    "scope_calibration": {
+      "title": "Scope calibration",
+      "status": "pass",
+      "detail": "claims bounded: can=2, cannot=2, exceeds_boundary=0",
+      "critical": false,
+      "blocks_high": true
+    },
+    "independent_study_count": {
+      "title": "Independent study-sample count",
+      "status": "pass",
+      "detail": "independent studies=6, samples=6",
+      "critical": true,
+      "blocks_high": true
+    },
+    "deterministic_confidence": {
+      "title": "Deterministic confidence",
+      "status": "pass",
+      "detail": "deterministic confidence=High (score=0.893, policy=2026-08-12.v3)",
+      "critical": true,
+      "blocks_high": true
+    },
+    "decision_action_consistency": {
+      "title": "Decision action consistency",
+      "status": "pass",
+      "detail": "ADOPT is supported: High confidence with direct primary-outcome evidence",
+      "critical": true,
+      "blocks_high": false
+    }
+  },
+  "passed": true,
+  "critical_failures": [],
+  "high_confidence_allowed": true,
+  "max_confidence": "High",
+  "enforcement": {
+    "rule": "confidence capped at {max}; High confidence requires a fully passing gate and >= 2 independent studies",
+    "max_confidence": "High",
+    "requires_action_change": false,
+    "action_override": null
+  }
+}

package/examples/spaced-retrieval-practice/methodology.json

@@ -0,0 +1,78 @@
+{
+  "target": "overall",
+  "verdict": "PASS",
+  "audit_items": {
+    "control_group": {
+      "status": "met",
+      "note": "由本轮真实检索到的证据集评估得出"
+    },
+    "randomization": {
+      "status": "partial",
+      "note": "由本轮真实检索到的证据集评估得出"
+    },
+    "pre_test": {
+      "status": "partial",
+      "note": "由本轮真实检索到的证据集评估得出"
+    },
+    "post_test": {
+      "status": "met",
+      "note": "由本轮真实检索到的证据集评估得出"
+    },
+    "retention_test": {
+      "status": "met",
+      "note": "由本轮真实检索到的证据集评估得出"
+    },
+    "transfer_test": {
+      "status": "partial",
+      "note": "由本轮真实检索到的证据集评估得出"
+    },
+    "sample_bias": {
+      "status": "partial",
+      "note": "由本轮真实检索到的证据集评估得出"
+    },
+    "self_selection": {
+      "status": "partial",
+      "note": "由本轮真实检索到的证据集评估得出"
+    },
+    "measurement_validity": {
+      "status": "met",
+      "note": "由本轮真实检索到的证据集评估得出"
+    },
+    "confounders": {
+      "status": "partial",
+      "note": "由本轮真实检索到的证据集评估得出"
+    },
+    "instructor_effect": {
+      "status": "missing",
+      "note": "由本轮真实检索到的证据集评估得出"
+    },
+    "novelty_effect": {
+      "status": "missing",
+      "note": "由本轮真实检索到的证据集评估得出"
+    },
+    "tool_version_effect": {
+      "status": "not_applicable",
+      "note": "由本轮真实检索到的证据集评估得出"
+    },
+    "ai_usage_policy": {
+      "status": "not_applicable",
+      "note": "由本轮真实检索到的证据集评估得出"
+    },
+    "dropout": {
+      "status": "missing",
+      "note": "由本轮真实检索到的证据集评估得出"
+    }
+  },
+  "task_vs_learning_guard": {
+    "measured_construct": "delayed retention and transfer, not immediate task speed",
+    "equates_task_with_learning": false,
+    "note": "本证据集以延迟保持与迁移为主要结果，未把即时练习表现等同于学习。"
+  },
+  "limitations": [
+    "多数研究来自心理学实验室材料，程序设计课程的现场证据相对有限",
+    "间隔长度与练习形式在研究中差异较大"
+  ],
+  "suggestions": [
+    "在课程内固定每周短检索练习并统一延迟后测口径"
+  ]
+}