{
  "name": "Comparison of ANOVA Variance-Explained Effect Size Statistics",
  "description": "Comparative compilation of variance-explained effect size statistics in ANOVA models, detailing mathematical denominators, additivity across model factors, sample estimator bias properties, Cohen benchmark thresholds, and cross-design comparability compiled from Cohen (1988), Olejnik and Algina (2003), Richardson (2011), and Lakens (2013).",
  "page": "https://casrai.org/guides/effect-size-eta-squared",
  "publisher": "CASRAI",
  "license": "https://creativecommons.org/licenses/by/4.0/",
  "dateModified": "2026-11-15T07:27:10Z",
  "temporalCoverage": null,
  "spatialCoverage": null,
  "columns": [
    {
      "name": "Statistic",
      "unit": "",
      "description": "Name and conventional mathematical symbol of the effect size metric"
    },
    {
      "name": "Denominator Formulation",
      "unit": "",
      "description": "Formulation of the variance denominator used to compute the effect size"
    },
    {
      "name": "Additive Property",
      "unit": "",
      "description": "Whether individual effect metrics sum additively to total model R-squared (bounded by 1.0)"
    },
    {
      "name": "Sample Estimator Bias",
      "unit": "",
      "description": "Directional bias of sample estimate relative to true population parameter"
    },
    {
      "name": "Cohen Small Cutoff",
      "unit": "proportion",
      "description": "Operational small effect size threshold per Cohen (1988)"
    },
    {
      "name": "Cohen Medium Cutoff",
      "unit": "proportion",
      "description": "Operational medium effect size threshold per Cohen (1988)"
    },
    {
      "name": "Cohen Large Cutoff",
      "unit": "proportion",
      "description": "Operational large effect size threshold per Cohen (1988)"
    },
    {
      "name": "Design Comparability",
      "unit": "",
      "description": "Applicability and validity of cross-study comparisons across varying research designs"
    }
  ],
  "rows": [
    {
      "Statistic": "Eta-Squared (η²)",
      "Denominator Formulation": "SS_total",
      "Additive Property": "Yes (sums to R² ≤ 1.0)",
      "Sample Estimator Bias": "Upwardly biased (overestimates population effect)",
      "Cohen Small Cutoff": 0.01,
      "Cohen Medium Cutoff": 0.06,
      "Cohen Large Cutoff": 0.14,
      "Design Comparability": "Between-subjects designs only"
    },
    {
      "Statistic": "Partial Eta-Squared (ηp²)",
      "Denominator Formulation": "SS_effect + SS_error",
      "Additive Property": "No (sum of effects can exceed 1.0)",
      "Sample Estimator Bias": "Upwardly biased (overestimates population effect)",
      "Cohen Small Cutoff": 0.01,
      "Cohen Medium Cutoff": 0.06,
      "Cohen Large Cutoff": 0.14,
      "Design Comparability": "Limited (design-dependent error structures)"
    },
    {
      "Statistic": "Generalized Eta-Squared (ηG²)",
      "Denominator Formulation": "SS_effect + SS_measured + SS_error",
      "Additive Property": "Design dependent",
      "Sample Estimator Bias": "Upwardly biased (overestimates population effect)",
      "Cohen Small Cutoff": 0.01,
      "Cohen Medium Cutoff": 0.06,
      "Cohen Large Cutoff": 0.14,
      "Design Comparability": "High (designed specifically for cross-design meta-analyses)"
    },
    {
      "Statistic": "Omega-Squared (ω²)",
      "Denominator Formulation": "SS_total + MS_error",
      "Additive Property": "Yes (sums to estimated population R²)",
      "Sample Estimator Bias": "Less biased (adjusted for degrees of freedom and error variance)",
      "Cohen Small Cutoff": 0.01,
      "Cohen Medium Cutoff": 0.06,
      "Cohen Large Cutoff": 0.14,
      "Design Comparability": "Between-subjects designs only"
    },
    {
      "Statistic": "Partial Omega-Squared (ωp²)",
      "Denominator Formulation": "SS_effect + (N - df_effect) * MS_error",
      "Additive Property": "No (non-additive across model factors)",
      "Sample Estimator Bias": "Less biased (adjusted for degrees of freedom and error variance)",
      "Cohen Small Cutoff": 0.01,
      "Cohen Medium Cutoff": 0.06,
      "Cohen Large Cutoff": 0.14,
      "Design Comparability": "Limited (cannot generalize across mixed designs)"
    },
    {
      "Statistic": "Epsilon-Squared (ε²)",
      "Denominator Formulation": "SS_total",
      "Additive Property": "Yes (sums to adjusted sample variance)",
      "Sample Estimator Bias": "Less biased (less sample bias than classical eta-squared)",
      "Cohen Small Cutoff": 0.01,
      "Cohen Medium Cutoff": 0.06,
      "Cohen Large Cutoff": 0.14,
      "Design Comparability": "Between-subjects designs only"
    }
  ],
  "sources": [
    {
      "title": "Statistical Power Analysis for the Behavioral Sciences",
      "url": "https://www.utstat.toronto.edu/~brunner/oldclass/378f16/readings/CohenPower.pdf",
      "accessed": "2026-11-15"
    },
    {
      "title": "Calculating and reporting effect sizes to facilitate cumulative science",
      "url": "https://doi.org/10.3389/fpsyg.2013.00863",
      "accessed": "2026-11-15"
    },
    {
      "title": "Generalized eta and omega squared statistics",
      "url": "https://doi.org/10.1037/1082-989X.8.4.434",
      "accessed": "2026-11-15"
    },
    {
      "title": "Eta squared and partial eta squared as measures of effect size in educational research",
      "url": "https://doi.org/10.1016/j.edurev.2010.12.001",
      "accessed": "2026-11-15"
    }
  ]
}