{
  "name": "Reported diagnostics and thresholds for inverse probability weighting",
  "description": "Compiled from a methods tutorial (Chesnaye et al. 2022), a methodological paper on weight construction (Cole and Hernan 2008) and the WeightIt R package vignette: the numeric checks and reference values each source gives for judging an IPW analysis.",
  "page": "https://casrai.org/guides/inverse-probability-weighting",
  "publisher": "CASRAI",
  "license": "https://creativecommons.org/licenses/by/4.0/",
  "dateModified": "2026-11-18T08:14:32Z",
  "temporalCoverage": "2008/2022",
  "spatialCoverage": null,
  "columns": [
    {
      "name": "Diagnostic or reference value",
      "unit": "",
      "description": ""
    },
    {
      "name": "Value",
      "unit": "",
      "description": "Value as reported by the source; 0.10 is a proportion"
    },
    {
      "name": "What it checks",
      "unit": "",
      "description": ""
    },
    {
      "name": "Source",
      "unit": "",
      "description": ""
    }
  ],
  "rows": [
    {
      "Diagnostic or reference value": "Standardized difference after weighting, negligible imbalance",
      "Value": "below 0.10 (10%)",
      "What it checks": "Covariate balance between weighted groups",
      "Source": "Chesnaye et al. 2022"
    },
    {
      "Diagnostic or reference value": "Example balance threshold in package output (bal.tab thresholds argument)",
      "Value": "m = 0.05",
      "What it checks": "Flags weighted mean differences above the threshold",
      "Source": "WeightIt vignette"
    },
    {
      "Diagnostic or reference value": "Mean of stabilized weights",
      "Value": "1",
      "What it checks": "Necessary condition for a correctly specified weight model",
      "Source": "Cole and Hernan 2008"
    },
    {
      "Diagnostic or reference value": "Common truncation percentiles",
      "Value": "1st and 99th",
      "What it checks": "Caps extreme weights; reduces variance, increases bias",
      "Source": "Chesnaye et al. 2022; Cole and Hernan 2008"
    },
    {
      "Diagnostic or reference value": "Cole and Hernan example: stabilized weight range after 1st/99th truncation",
      "Value": "0.20 to 4.49",
      "What it checks": "Spread of weights in their worked example",
      "Source": "Cole and Hernan 2008"
    },
    {
      "Diagnostic or reference value": "Cole and Hernan example: estimate, untruncated (SE)",
      "Value": "-1.91 (0.132)",
      "What it checks": "Effect estimate before truncation",
      "Source": "Cole and Hernan 2008"
    },
    {
      "Diagnostic or reference value": "Cole and Hernan example: estimate, truncated at 1st/99th (SE)",
      "Value": "-1.80 (0.122)",
      "What it checks": "Effect estimate after truncation",
      "Source": "Cole and Hernan 2008"
    },
    {
      "Diagnostic or reference value": "Sample size below which variance underestimation is a concern",
      "Value": "150",
      "What it checks": "Small-sample reliability of IPTW variance",
      "Source": "Chesnaye et al. 2022"
    }
  ],
  "sources": [
    {
      "title": "Chesnaye NC et al. An introduction to inverse probability of treatment weighting in observational research. Clin Kidney J 2022",
      "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC8757413",
      "accessed": "2026-11-18"
    },
    {
      "title": "Cole SR, Hernan MA. Constructing inverse probability weights for marginal structural models. Am J Epidemiol 2008;168(6):656-664",
      "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC2732954/",
      "accessed": "2026-11-18"
    },
    {
      "title": "WeightIt: Weighting for covariate balance (CRAN vignette)",
      "url": "https://cran.r-project.org/web/packages/WeightIt/vignettes/WeightIt.html",
      "accessed": "2026-11-18"
    }
  ]
}