[
  {
    "type": "Observed",
    "item": "Collision numerator",
    "detail": "Published injury-collision counts, road location. Damage-only incidents and insurance claim costs are absent."
  },
  {
    "type": "Estimated",
    "item": "Mileage denominator",
    "detail": "DfT estimated all-motor vehicle miles, joined by authority and year. Realised target-year exposure is used for assessment; this is not an at-origin mileage forecast."
  },
  {
    "type": "Constant",
    "item": "Census context",
    "detail": "2011 age, density and qualification shares are held fixed through 2018–2025. They describe residents around roads, not drivers."
  },
  {
    "type": "Observed",
    "item": "Historical deprivation",
    "detail": "Use 2015/2019 published income and employment scores by recorded availability. No 2025 deprivation or 2021 education backdating."
  },
  {
    "type": "Allocated",
    "item": "Mileage stress scenarios only",
    "detail": "Inner/Outer reweighting is a hypothetical authority-level allocation; two scenarios preserve annual London totals. No LSOA mileage is allocated in the main study."
  },
  {
    "type": "Estimated",
    "item": "Model and selection",
    "detail": "NB2 fits, training-only history prior and transformations. The 1% deviance selection tolerance favours stronger pooling. Models and gates are retained unchanged."
  },
  {
    "type": "Constant",
    "item": "Frozen cohort and folds",
    "detail": "133 complete-history English authorities; 33 London boroughs assessed. 2020/2021 are excluded as training outcomes but can appear as lagged predictors. Cohort inclusion uses retrospective data availability."
  },
  {
    "type": "Constant",
    "item": "Uncertainty design",
    "detail": "399 authority-history bootstrap draws; four failures retained, 395 complete. Marginal percentile ranges omit residual spatial dependence and multiplicity adjustment."
  }
]