Claim at v1.2.0

moon-full-emergencies

a9cba31063017272d633ccc6af3d2b5f238d186001a12792da1e8c2c56c1a730

Data

{
  "has_outcome": true,
  "kind": "confirmation",
  "lineage": {
    "audit": "docs/lineage_audit.md; data/derived/lineage_audit.json",
    "class": "none",
    "class_meaning": "the audit found no city-data source fetched before registration",
    "classes": [],
    "early_sources": []
  },
  "outcome": "moon-full-emergencies",
  "prior": 0.6,
  "registered_at": "2026-11-14T07:10:45-0500",
  "registrant": "Claude (coordinator), for Jeff Vernon",
  "registration": {
    "already_seen": "The three daily series were built for earlier claims. None has been tabulated or plotted against lunar phase by this registrant. The equivalence bounds were set from expected standard errors, not from results.",
    "alternatives": [
      "Moonlight changes visibility and night-time activity (a real but non-lunacy path)",
      "Cloud cover hides the moon and dilutes any effect",
      "Chance among many lunar studies (publication of positive results)"
    ],
    "can_it_fail": "Yes. The graph's crash finding (1.042) lies outside the crash bound, so a New York effect of that size would miss support_if, and its interval would probably exclude 1.00. Expected standard errors: about 0.025 for shootings, 0.004 for ambulance incidents and 0.008 for injury crashes.",
    "graph_evidence": [
      {
        "association_id": "variable search 'full moon', 'lunar cycle', 'moon phase', 'lunar phase' (14 November 2026)",
        "concepts": [
          "full moon",
          "road traffic crash emergency transport"
        ],
        "direction": "+",
        "n_findings": 1,
        "n_studies": 1,
        "pubmed_ids": [
          29990928
        ],
        "summary": "Relative risk 1.042 (1.021 to 1.063) of emergency transport for road traffic crashes on full-moon nights."
      },
      {
        "association_id": "same search",
        "concepts": [
          "full moon",
          "motorcycle deaths"
        ],
        "direction": "+",
        "n_findings": 2,
        "n_studies": 1,
        "pubmed_ids": [
          29229755
        ],
        "summary": "Relative risk 1.05 (1.02 to 1.09) of motorcycle deaths on full-moon nights; a second record gives an odds ratio 1.26."
      },
      {
        "association_id": "same search",
        "concepts": [
          "moon phase",
          "homicide"
        ],
        "direction": "-",
        "n_findings": 2,
        "n_studies": 1,
        "pubmed_ids": [
          30647032
        ],
        "summary": "Homicides in Finland: relative risk 0.85 (0.77 to 0.94) at full moon."
      },
      {
        "association_id": "same search",
        "concepts": [
          "full moon",
          "out-of-hospital cardiac arrest"
        ],
        "direction": "0",
        "n_findings": 3,
        "n_studies": 1,
        "pubmed_ids": [
          33693140
        ],
        "summary": "Relative risk 1.007 (0.993 to 1.021)."
      }
    ],
    "graph_evidence_note": "The variable search returned 60 findings from 36 studies on lunar variables. The values above are the graph's records as returned by the API; the abstracts were not read before registration.",
    "id": "moon-full-emergencies",
    "kind": "confirmation",
    "literature": [
      {
        "citation": "Rotton J, Kelly IW. Much ado about the full moon: a meta-analysis of lunar-lunacy research. Psychol Bull 1985;97(2):286-306.",
        "doi": "10.1037/0033-2909.97.2.286",
        "finding": "Phases of the moon accounted for no more than about 1% of the variance in the behaviours studied (cited from memory; to be verified)."
      }
    ],
    "plan": "analyses/moon/test.py, committed with this registration and not yet run. Poisson regression of daily counts on a 3-day full-moon window (PyEphem instants, New York dates), with day of week, year-by-month fixed effects, holiday, daily maximum temperature and its square, and a wet-day indicator; Newey-West errors, 7 lags. Outcomes: shooting incidents 2006-2025 (data/derived/heat/shootings_daily.csv), all EMS incidents 2017-01-01 to 2026-08-31 (data/derived/health_gaps/ems_daily.csv), crashes with an injured person 2014-2025 (Motor Vehicle Collisions, h9gi-nx95, the copy fetched for the turn-restriction claims). Placebo (rule 12): the same model with a 3-day window that recurs every 23 days. Secondary, deciding nothing: the full-moon day alone.",
    "prediction": {
      "direction": "0",
      "magnitude": "Rate ratio for the 3-day full-moon window: shootings 0.96 to 1.04, ambulance incidents 0.995 to 1.005, injury crashes 0.98 to 1.02. Best guess 1.00 for each.",
      "refute_if": "For at least one outcome the 98.3% interval excludes 1.00 while the 95% interval of the 23-day placebo for that outcome includes 1.00.",
      "support_if": "All three 95% intervals lie inside their equivalence bounds: shootings 0.93 to 1.07, ambulance incidents 0.98 to 1.02, injury crashes 0.97 to 1.03."
    },
    "prior": 0.6,
    "prior_note": "The ambulance bound is narrow; a small real difference or residual seasonality could miss it.",
    "registered_at": "2026-11-14T07:10:45-0500",
    "registrant": "Claude (coordinator), for Jeff Vernon",
    "sha256": "e9097b4a39e764212a31c6646953e0cf2d31fdeae07451bebe7dea84a2c598bb",
    "statement": "On the day of a full moon and the days either side of it, New York City has no more and no fewer shootings, ambulance incidents or injury crashes than on other days, within stated equivalence bounds."
  },
  "registration_commit": {
    "commit": "f673f98dcd0cf1d6150dcf1e58be357dd3a2d6b0",
    "committed_at": "2026-11-14T07:10:45-05:00"
  },
  "registration_file_sha256": "e0f412aa673e9062eb51427a2dcfc710f1c9e07a3c838308c4dcb26ba631ed0b",
  "registration_path": "claims/registered/moon-full-emergencies.json",
  "registration_sha256": "e9097b4a39e764212a31c6646953e0cf2d31fdeae07451bebe7dea84a2c598bb",
  "registration_text": "{\n \"id\": \"moon-full-emergencies\",\n \"kind\": \"confirmation\",\n \"statement\": \"On the day of a full moon and the days either side of it, New York City has no more and no fewer shootings, ambulance incidents or injury crashes than on other days, within stated equivalence bounds.\",\n \"graph_evidence\": [\n  {\n   \"association_id\": \"variable search 'full moon', 'lunar cycle', 'moon phase', 'lunar phase' (14 November 2026)\",\n   \"concepts\": [\n    \"full moon\",\n    \"road traffic crash emergency transport\"\n   ],\n   \"n_findings\": 1,\n   \"n_studies\": 1,\n   \"direction\": \"+\",\n   \"summary\": \"Relative risk 1.042 (1.021 to 1.063) of emergency transport for road traffic crashes on full-moon nights.\",\n   \"pubmed_ids\": [\n    29990928\n   ]\n  },\n  {\n   \"association_id\": \"same search\",\n   \"concepts\": [\n    \"full moon\",\n    \"motorcycle deaths\"\n   ],\n   \"n_findings\": 2,\n   \"n_studies\": 1,\n   \"direction\": \"+\",\n   \"summary\": \"Relative risk 1.05 (1.02 to 1.09) of motorcycle deaths on full-moon nights; a second record gives an odds ratio 1.26.\",\n   \"pubmed_ids\": [\n    29229755\n   ]\n  },\n  {\n   \"association_id\": \"same search\",\n   \"concepts\": [\n    \"moon phase\",\n    \"homicide\"\n   ],\n   \"n_findings\": 2,\n   \"n_studies\": 1,\n   \"direction\": \"-\",\n   \"summary\": \"Homicides in Finland: relative risk 0.85 (0.77 to 0.94) at full moon.\",\n   \"pubmed_ids\": [\n    30647032\n   ]\n  },\n  {\n   \"association_id\": \"same search\",\n   \"concepts\": [\n    \"full moon\",\n    \"out-of-hospital cardiac arrest\"\n   ],\n   \"n_findings\": 3,\n   \"n_studies\": 1,\n   \"direction\": \"0\",\n   \"summary\": \"Relative risk 1.007 (0.993 to 1.021).\",\n   \"pubmed_ids\": [\n    33693140\n   ]\n  }\n ],\n \"graph_evidence_note\": \"The variable search returned 60 findings from 36 studies on lunar variables. The values above are the graph's records as returned by the API; the abstracts were not read before registration.\",\n \"literature\": [\n  {\n   \"citation\": \"Rotton J, Kelly IW. Much ado about the full moon: a meta-analysis of lunar-lunacy research. Psychol Bull 1985;97(2):286-306.\",\n   \"doi\": \"10.1037/0033-2909.97.2.286\",\n   \"finding\": \"Phases of the moon accounted for no more than about 1% of the variance in the behaviours studied (cited from memory; to be verified).\"\n  }\n ],\n \"prediction\": {\n  \"direction\": \"0\",\n  \"magnitude\": \"Rate ratio for the 3-day full-moon window: shootings 0.96 to 1.04, ambulance incidents 0.995 to 1.005, injury crashes 0.98 to 1.02. Best guess 1.00 for each.\",\n  \"support_if\": \"All three 95% intervals lie inside their equivalence bounds: shootings 0.93 to 1.07, ambulance incidents 0.98 to 1.02, injury crashes 0.97 to 1.03.\",\n  \"refute_if\": \"For at least one outcome the 98.3% interval excludes 1.00 while the 95% interval of the 23-day placebo for that outcome includes 1.00.\"\n },\n \"plan\": \"analyses/moon/test.py, committed with this registration and not yet run. Poisson regression of daily counts on a 3-day full-moon window (PyEphem instants, New York dates), with day of week, year-by-month fixed effects, holiday, daily maximum temperature and its square, and a wet-day indicator; Newey-West errors, 7 lags. Outcomes: shooting incidents 2006-2025 (data/derived/heat/shootings_daily.csv), all EMS incidents 2017-01-01 to 2026-08-31 (data/derived/health_gaps/ems_daily.csv), crashes with an injured person 2014-2025 (Motor Vehicle Collisions, h9gi-nx95, the copy fetched for the turn-restriction claims). Placebo (rule 12): the same model with a 3-day window that recurs every 23 days. Secondary, deciding nothing: the full-moon day alone.\",\n \"can_it_fail\": \"Yes. The graph's crash finding (1.042) lies outside the crash bound, so a New York effect of that size would miss support_if, and its interval would probably exclude 1.00. Expected standard errors: about 0.025 for shootings, 0.004 for ambulance incidents and 0.008 for injury crashes.\",\n \"already_seen\": \"The three daily series were built for earlier claims. None has been tabulated or plotted against lunar phase by this registrant. The equivalence bounds were set from expected standard errors, not from results.\",\n \"alternatives\": [\n  \"Moonlight changes visibility and night-time activity (a real but non-lunacy path)\",\n  \"Cloud cover hides the moon and dilutes any effect\",\n  \"Chance among many lunar studies (publication of positive results)\"\n ],\n \"prior\": 0.6,\n \"prior_note\": \"The ambulance bound is narrow; a small real difference or residual seasonality could miss it.\",\n \"registrant\": \"Claude (coordinator), for Jeff Vernon\",\n \"registered_at\": \"2026-11-14T07:10:45-0500\",\n \"sha256\": \"e9097b4a39e764212a31c6646953e0cf2d31fdeae07451bebe7dea84a2c598bb\"\n}",
  "statement": "On the day of a full moon and the days either side of it, New York City has no more and no fewer shootings, ambulance incidents or injury crashes than on other days, within stated equivalence bounds."
}

History in this collection

  1. addedv1.1.0Nov 14, 2026this version