Prior distributions for variance parameters in hierarchical models (comment on article by Browne and Draper)
Open accessThe DOI matches its Crossref record; 42 of the 42 references it cites were found in a registry or an index; 1 of the 23 sources its claims cite were checked for support.
Abstract
Various noninformative prior distributions have been suggested for scale parameters in hierarchical models. We construct a new folded-noncentral-$t$ family of conditionally conjugate priors for hierarchical standard deviation parameters, and then consider noninformative and weakly informative priors in this family. We use an example to illustrate serious problems with the inverse-gamma family of "noninformative" prior distributions. We suggest instead to use a uniform prior on the hierarchical standard deviation, using the half-$t$ family when the number of groups is small and in other settings where a weakly informative prior is desired. We also illustrate the use of the half-$t$ family for hierarchical modeling of multiple variance parameters such as arise in the analysis of variance.
Overview
CoChat Summary
TL;DR
Gelman argues that inverse-gamma “noninformative” priors for hierarchical variance parameters can badly distort posterior inference, especially when the group-level variance is near zero or the number of groups is small. He recommends a uniform prior on the hierarchical standard deviation when groups are not few, and half-t/half-Cauchy weakly informative priors otherwise, developed within a folded-noncentral-t conditionally conjugate family.
Key findings
- The inverse-gamma(ε, ε) family is not reliably noninformative: as ε → 0 it yields an improper posterior, and in the 8-schools example inverse-gamma(1, 1) and inverse-gamma(0.001, 0.001) priors on σ²_α constrained posterior inferences and increased shrinkage relative to a uniform prior on σ_α.
- A uniform prior on σ_α performed well for the 8-schools data (J = 8), but with only J = 3 schools it produced an extremely long right tail and undershrinkage; a half-Cauchy prior with scale A = 25 removed much of that unrealistic upper tail.
- For multiple variance components, Gelman recommends a hierarchical half-Cauchy model with scale estimated from data; in the 5 × 5 × 2 split-plot latin square Anova with nine variance components, the Cauchy scale parameter A was estimated at 1.8 with a 95% posterior interval of [0.5, 5.1].
- The folded-noncentral-t family, obtained via a redundant multiplicative parameterization, provides conditionally conjugate priors for σ_α and includes the improper uniform density (ν = −1) and proper half-Cauchy (ν = 1) as special cases.
Methods
- Conceptual/comment paper using a simple two-level normal hierarchical model: y_ij ∼ N(μ + α_j, σ²_y), α_j ∼ N(0, σ²_α), focusing on the prior for σ_α.
- Illustrates with the 8-schools example (J = 8, 6000 Bugs simulation draws), a 3-schools subset, and a 5 × 5 × 2 split-plot latin square Anova with nine variance components.
- Uses a parameter-expanded form y_ij ∼ N(μ + ξη_j, σ²_y), η_j ∼ N(0, σ²_η), with σ_α = |ξ|σ_η, to construct and compare uniform, inverse-gamma, half-Cauchy, and hierarchical half-Cauchy priors.
Limitations
- The paper states that noninformative and weakly informative priors are inherently provisional: after fitting, one should inspect the posterior and revise the prior if it does not make sense.
- The paper notes that the uniform prior on σ_α has a positive miscalibration and yields an improper posterior for J = 1 or 2; inverse-gamma inferences are sensitive to ε; and half-Cauchy scale choices are weakly informative rather than fully objective.
- Evident from the text: the recommendations rest on illustrative examples and conceptual argument rather than a broad simulation study or systematic empirical comparison, and extensions to nonlinear, multivariate, and more complex hierarchical models are left for further work.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
0 Valid · 1 Not valid · 14 Not checked · 56 Own results and interpretations
The cited sources that could be checked couldn’t confirm or dispute this claim.
In the paper
“Similarly, there are no always-nonnegative classical unbiased estimators of σ α or σ 2 α in the hierarchical model.”
- Also stated in: § 7.1 Prior distributions for variance parameters · plain
“Similarly, there are no alwaysnonnegative classical unbiased estimators of σ α or σ 2 α in the hierarchical model.”
§ Prior distributions for variance parameters in hierarchical models(Comment on Article by Browne and Draper) > 2.4 Calibration
What the sources say
- Could not verifycites Bickel and Blackwell (1967) matched to “A Note on Bayes Estimates” (1967) sourceExperimental — accuracy not yet measured
The introduction (p. 1) and Theorem 1.3 (p. 3) concern when an unbiased Bayes estimate must have zero Bayes risk. Searches for hierarchical models, σₐ, σ²ₐ, and always-nonnegative classical unbiased estimators found no passage addressing the claimed nonexistence result. (LLM, conf=0.77)
- Not checked — no full text of the sourcecites Meng and Zaslavsky (2002) matched to “Single observation unbiased priors” (2002) source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Also stated in: § 7.1 Prior distributions for variance parameters · plain
A result the paper states — it cites no source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“If J is low, the uniform prior density tends to lead to high estimates of σ α , as discussed in Section 5.2.”
§ Prior distributions for variance parameters in hierarchical models(Comment on Article by Browne and Draper) > 7.1 Prior distributions for variance parameters
hedgedstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“In contrast, the hierarchical half-Cauchy model performs a great deal of shrinkage, especially of the high ranges of the intervals.”
- Also stated in: § Bugs model: a half-Cauchy prior distribution on sigma.theta is induced # using a normal prior on xi and an inverse-gamma on tau.eta model { for (j in 1:J){ # J = the number of schools y[j] ~dnorm (theta[j], tau.y[j]) # data model: the likelihood theta[j] <-mu.theta+ xi*eta[j] tau.y[j] <-pow(sigma.y[j],-2) } xi ~dnorm (0, tau.xi) tau.xi <-pow(prior.scale,-2) for (j in 1:J){ eta[j] ~dnorm (0, tau.eta) # hierarchical model for theta } tau.eta ~dgamma (.5, .5)# chi^2 with 1 d.f.sigma.theta<-abs(xi)/sqrt(tau.eta)# cauchy = normal/sqrt(chi^2) mu.theta ~dnorm (0.0, 1.0E-6) # noninformative prior on mu } · plain
“The half-Cauchy model gives much sharper inferences, using the partial pooling that comes with fitting a hierarchical model.”
§ Prior distributions for variance parameters in hierarchical models(Comment on Article by Browne and Draper) > 6.2 Superpopulation and finite-population standard deviations
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Also stated in: § Bugs model: a half-Cauchy prior distribution on sigma.theta is induced # using a normal prior on xi and an inverse-gamma on tau.eta model { for (j in 1:J){ # J = the number of schools y[j] ~dnorm (theta[j], tau.y[j]) # data model: the likelihood theta[j] <-mu.theta+ xi*eta[j] tau.y[j] <-pow(sigma.y[j],-2) } xi ~dnorm (0, tau.xi) tau.xi <-pow(prior.scale,-2) for (j in 1:J){ eta[j] ~dnorm (0, tau.eta) # hierarchical model for theta } tau.eta ~dgamma (.5, .5)# chi^2 with 1 d.f.sigma.theta<-abs(xi)/sqrt(tau.eta)# cauchy = normal/sqrt(chi^2) mu.theta ~dnorm (0.0, 1.0E-6) # noninformative prior on mu } · plain
A result the paper states — it cites no source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“This prior distribution is even more sharply peaked near zero and further distorts posterior inferences, with the problem arising because the marginal likelihood for σ α remains high near zero.”
§ Prior distributions for variance parameters in hierarchical models(Comment on Article by Browne and Draper) > 5.1 Noninformative prior distributions for the 8-schools problem
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The uniform prior distribution seems fine for the 8-school analysis”
- Also stated in: § 5.2 Weakly informative prior distribution for the 3-schools problem · plain
“the default uniform prior gave reasonable results”
- Also stated in: § 7.1 Prior distributions for variance parameters · hedged
“Figure 1 illustrates the generally robust properties of the uniform prior density on σ α .”
§ Prior distributions for variance parameters in hierarchical models(Comment on Article by Browne and Draper) > 5.2 Weakly informative prior distribution for the 3-schools problem
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Also stated in: § 5.2 Weakly informative prior distribution for the 3-schools problem · plain
26 more claims
Create a free account to see all 71 claims and every source verdict.
- It may state two separate claims as one. 34
- Probably not a substantial claim. 12
- The quote only partly states this claim. 1
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Because the uniform distribution tends to unduly favor presence of heterogeneity in discrimination and calibration estimates across studies,58,59”
From A framework for meta-analysis of prediction model studies with binary and time-to-event outcomes (2019) · cites this paper as [59]
Experimental — accuracy not yet measured
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Prior distributions for variance parameters in hierarchical models(Comment on Article by Browne and Draper) > 7.1 Prior distributions for variance parameters
“we see conditional conjugacy in the wider class of half-t distributions on σ α , which include the uniform and half-Cauchy densities on σ α (as well as inverse-gamma on σ 2 α ) as special cases.”
§ Prior distributions for variance parameters in hierarchical models(Comment on Article by Browne and Draper) > 3.2 Folded-noncentral-t prior distribution for σ α
“In the limit that the denominator is specified exactly, we have a folded normal distribution; conversely, specifying the numerator exactly yields the square-root-inverse-χ 2 distribution for σ α , as in Section 3.1.”
Citations
Cites
Gelman, A., Carlin, J. B., Stern, H. S., & Rubin, D. B. (2003) · doi:10.1201/9780429258480 · Cited by 27,276
Self-citationKass, R. E., & Raftery, A. E. (1995) · Journal of the American Statistical Association · doi:10.1080/01621459.1995.10476572 · Cited by 12,558
Sampling-based approaches to calculating marginal densities
Gelfand, A. E., & Smith, A. F. M. (1990) · Journal of the American Statistical Association · doi:10.1080/01621459.1990.10476213 · Cited by 6,606 · cited 1×
“Typically, enough data will be available to estimate µ and σy that one can use any reasonable noninformative prior distribution—for example, p(µ, σy) ∝ 1 or p(µ, log σy) ∝ 1.”
Introducing Multilevel Modeling
Kreft, I., & de Leeuw, J. (1998) · doi:10.4135/9781849209366 · Cited by 2,997
POSTERIOR PREDICTIVE ASSESSMENT OF MODEL FITNESS VIA REALIZED DISCREPANCIES
Gelman, A., Meng, X., & Stern, H. S. (1996) · Cited by 2,638
Self-citationBayes and Empirical Bayes Methods for Data Analysis
(2001) · Technometrics · doi:10.1198/tech.2001.s608 · Cited by 2,098
INADMISSIBILITY OF THE USUAL ESTIMATOR FOR THE MEAN OF A MULTIVARIATE NORMAL DISTRIBUTION
Stein, C. (1956) · Contribution to the Theory of Statistics · doi:10.1525/9780520313880-018 · Cited by 1,841
Bayesian Approaches to Clinical Trials and Health‐Care Evaluation
Spiegelhalter, D. J., Abrams, K. R., & Myles, J. P. (2003) · Wiley eBooks · doi:10.1002/0470092602 · Cited by 1,673
Parametric empirical Bayes inference: theory and applications
Morris, C. N. (1983) · Journal of the American Statistical Association · doi:10.1080/01621459.1983.10477920 · Cited by 1,448 · cited 1×
“On the theoretical side, hierarchical models allow a more “objective” approach to inference by estimating the parameters of prior distributions from data rather than requiring them to be specified using subjective information (see James and Stein, 1960, Efron and Morris, 1975, and Morris, 1983).”
The Selection of Prior Distributions by Formal Rules
Kass, R. E., & Wasserman, L. (1996) · Journal of the American Statistical Association · doi:10.1080/01621459.1996.10477003 · Cited by 1,308
Research works printed but not linked: 5
- Box, G. E. P. and Tiao, G. C. (1973). Bayesian Inference in Statistical Analysis. Reading, Mass.: Addison-Wesley.
- James, W. and Stein, C. (1960). "Estimation with quadratic loss." In Neyman, J. (ed.), Proceedings of the Fourth Berkeley Symposium, volume 1, 361–380. Berkeley: University of California Press.
- Johnson, N. L. and Kotz, S. (1972). Distributions in Statistics. New York: Wiley. 4 vols.
- Liu, C. (2001). "Bayesian analysis of multivariate probit models. Discussion of “The art of data augmentation” by D. A. van Dyk and X. L. Meng." Journal of Computational and Graphical Statistics, 10: 75–81.
- Raudenbush, S. W. and Bryk, A. S. (2002). Hierarchical Linear Models. Thousand Oaks, Calif.: Sage., second edition.
Cited by
Items 3,101–3,200 of 4,089
Horseshoe‐based Bayesian nonparametric estimation of effective population size trajectories
Faulkner, J. R., Magee, A. F., Shapiro, B., & Minin, V. N. (2020) · Biometrics · doi:10.1111/biom.13276 · Cited by 25
Gerard, E., Zohar, S., Thai, H., Lorenzato, C., Riviere, M., & Ursino, M. (2021) · Biometrics · doi:10.1111/biom.13433 · Cited by 21
“Regarding the half-Cauchy distribution, we followed the recommendations by Gelman, as we assumed that τz could be near 0 [10].”
MethodA Method to Assess and Argue for Practical Significance in Software Engineering
Torkar, R., Furia, C. A., FELDT, R. H., Neto, F. G. D. O., Gren, L., Lenberg, P., & Ernst, N. (2021) · IEEE Transactions on Software Engineering · doi:10.1109/tse.2020.3048991 · Cited by 28
Campos, L. F., Glickman, M. E., & Hunter, K. B. (2019) · Biostatistics · doi:10.1093/biostatistics/kxz059 · Cited by 1
State space modeling for analysis of behavior in learning experiments
Smith, A. C. (2015) · Cambridge University Press eBooks · doi:10.1017/cbo9781139941433.011 · Cited by 2
Asar, Ö. (2020) · Statistical Modelling · doi:10.1177/1471082x20920122 · Cited by 1
“Standard deviations of the distributions, e.g. for √ π2 3 (φ−2 − 1) for Bridge, are assigned half-Cauchy with location 0 and scale 5 (Gelman, 2006; Polson and Scott, 2012).”
BackgroundBayesian simultaneous prediction intervals and bounds for a finite population
Hamada, M. S., & Weaver, B. (2019) · Quality Engineering · doi:10.1080/08982112.2019.1656811 · Cited by 1
Parthasarathy, B., Somanathan, H., & WRIGHT, J. M. (2021) · Ethology · doi:10.1111/eth.13229 · Cited by 1
“We used parameter extended priors (Gelman, 2006) with the prior means (alpha.mu) set to 0 and the prior covariance (alpha.”
MethodWolf-Gonzalez, G., Murphy, S. M., Springer, M. T., & Cox, J. J. (2022) · Journal of Arid Environments · doi:10.1016/j.jaridenv.2022.104793 · Cited by 1
Bayesian borrowing from historical control data in a vaccine efficacy trial
Peng, L., Jin, J., Chambonneau, L., Zhao, X., & Zhang, J. (2023) · Pharmaceutical Statistics · doi:10.1002/pst.2313 · Cited by 1
Odur, B., Nansubuga, E., Wamala, R., et al. (2023) · East African Journal of Health and Science · doi:10.37284/eajhs.6.1.1346 · Cited by 1
“Traditional linear regression techniques are not capable of handling data with a hierarchical multilevel structure since data with a hierarchical structure violates an important assumption of the linear regression model (Gelman, 2006).”
BackgroundBayesian hierarchical model for dose-finding trial incorporating historical data
Han, L., Deng, Q., He, Z., Fleischer, F., & Yu, F. (2023) · Journal of Biopharmaceutical Statistics · doi:10.1080/10543406.2023.2251578 · Cited by 2
“This prior distribution should cover the range of typical values representing plausible (e.g., from small to large) between-trial heterogeneity (Gelman 2006; Röver et al. 2021; Spiegelhalter et al. 2004).”
Estimating Commercial Property Fundamentals from REIT data
Geltner, D., Kumar, A., & de Minne, A. V. (2023) · The Journal of Real Estate Finance and Economics · doi:10.1007/s11146-023-09962-z · Cited by 1
“N (0,10)), whereas the other parameters are largely uninformative ( N (0,1)) (Gelman, 2006).”
Bayesian Semiparametric Local Clustering of Multiple Time Series Data
Fan, J., & Sarkar, A. (2023) · Technometrics · doi:10.1080/00401706.2023.2288324 · Cited by 2
Fu, H., Tang, L., Rosen, O., Hipwell, A. E., Huppert, T. J., & Krafty, R. T. (2023) · Biostatistics · doi:10.1093/biostatistics/kxad034 · Cited by 1
Alpine shrub leaf litter decomposition across mountain summits in south‐eastern Australia
Venn, S., Camac, J., Grover, S., et al. (2024) · Austral Ecology · doi:10.1111/aec.13511 · Cited by 1
Wedel, M., Chen, D., & Kopyakova, A. (2024) · Journal of Marketing Analytics · doi:10.1057/s41270-024-00312-3 · Cited by 1
“The use of default half-Cauchy priors has been advocated on theoretical grounds (Gelman 2006; McNeish 2016; Pol-son and Scott 2012) and these priors have empirically been shown to perform well (McNeish and Stapleton 2016).”
Lu, Y., Zheng, Q., & Henning, K. S. S. (2024) · Multivariate Behavioral Research · doi:10.1080/00273171.2024.2358233 · Cited by 1
“A line of literature has suggested the use of the half-Cauchy distribution (as a special case of the folded noncentral t -distribution) as a default prior for a hierarchical standard deviation parameter, e.g., Gelman (2006) and Polson and Scott (2012).”
Approximate Cross-Validated Mean Estimates for Bayesian Hierarchical Regression Models
Zhang, A., Daniels, M. J., Li, C., & Bao, L. (2024) · Journal of Computational and Graphical Statistics · doi:10.1080/10618600.2024.2404711 · Cited by 4
“There is also a growing preference for half-t priors over the conjugate inverse-gamma (Gelman 2006; Polson and Scott 2012).”
Lo, M. K., Leung, Y. F., Chan, C. L., et al. (2024) · Canadian Geotechnical Journal · doi:10.1139/cgj-2023-0633 · Cited by 1
Regularized Variational Approximation for Partially Confirmatory Factor Analysis
Jin, Y., & Chen, J. (2024) · Structural Equation Modeling A Multidisciplinary Journal · doi:10.1080/10705511.2024.2432612 · Cited by 4
A Dynamic Micro Panel Data Model on Individual Expenditure Habit: Hierarchical Bayesian Method
Omolara, A. Y., Omowumi, A. N., & Moromoke, M. H. (2025) · International Journal of Scientific Research in Science and Technology · doi:10.32628/ijsrst24114324 · Cited by 1
Reschützegger, T., & Salau, N. P. G. (2025) · Cleaner Chemical Engineering · doi:10.1016/j.clce.2025.100146 · Cited by 3
Zillig, K. W., Bell, H. N., FitzGerald, A. M., & Fangue, N. A. (2025) · Frontiers in Fish Science · doi:10.3389/frish.2024.1508746 · Cited by 3
Congdon, P. D. (2024) · Journal of the Royal Statistical Society Series A (Statistics in Society) · doi:10.1093/jrsssa/qnae156 · Cited by 1
Recurrent events modeling based on a reflected Brownian motion with application to hypoglycemia
Xie, Y., Fu, H., Huang, Y., Pozdnyakov, V. A., & Yan, J. (2024) · Biostatistics · doi:10.1093/biostatistics/kxae053 · Cited by 2
“The hyper-parameters ( a, b ) of the inverse gamma prior for the frailty variances are often set to be a = b = 1 (Gelman, 2006; Manda and Meyer, 2005).”
BackgroundMenzies, N. A. (2025) · Populations · doi:10.3390/populations1010003 · Cited by 1
“In general, weakly informative priors [26] were specified, except where substantial prior information was available, such as with age-specific population mortality rates.”
Estimation of Pressure Leakage Severity for Space Habitats Using Extended Kalman Filter
Mirfarah, M., Lund, A., & Dyke, S. J. (2025) · AIAA Journal · doi:10.2514/1.j063825 · Cited by 1
Active learning for regression in engineering populations: a risk-informed approach
Clarkson, D. R., Bull, L. A., DARDENO, T. A., Wickramarachchi, C. T., Cross, E. J., Rogers, T. J., Worden, K., Dervilis, N., & Hughes, A. J. (2025) · Data-Centric Engineering · doi:10.1017/dce.2025.7 · Cited by 3
“As recommended by Gelman (2006), the variance priors for this hierarchical model are set to be weakly informative.”
FRODO: a novel approach to micro–macro multilevel regression
McDonald, S., Leblanc, A., Muthukumarana, S., & Campbell, D. (2025) · Computational Statistics · doi:10.1007/s00180-025-01631-4 · Cited by 1
Muell, M. R., Hall, J. M., Smith, K. V., Oaks, J. R., Wolak, M. E., & Warner, D. A. (2025) · Evolution · doi:10.1093/evolut/qpaf104 · Cited by 2
Zhang, J., Zhou, H., Wages, N. A., Guo, Z., Liu, F., Jemielita, T., Yan, F., & Lin, R. (2025) · Statistics in Medicine · doi:10.1002/sim.70090 · Cited by 1
Honey, I shrunk the irrelevant effects! Simple and flexible approximate Bayesian regularization
Karimova, D., van Erp, S., Leenders, R., & Mulder, J. (2025) · Journal of Mathematical Psychology · doi:10.1016/j.jmp.2025.102925 · Cited by 1
Assessing predictive probability of success for future clinical trials
Sachdeva, A., Tiwari, R. C., & Zhou, M. (2025) · Journal of Biopharmaceutical Statistics · doi:10.1080/10543406.2025.2510262 · Cited by 1
Rajeb, M., Ma, W., He, Q., et al. (2025) · Journal of Educational and Behavioral Statistics · doi:10.3102/10769986251334788 · Cited by 1
Yamaguchi, Y., Takeda, K., Takeda, J., & Maruo, K. (2025) · Statistics in Biopharmaceutical Research · doi:10.1080/19466315.2025.2540541 · Cited by 1
Accounting for correlation and censoring in Bayesian Network Scale-up Method Models
Vogel, B., Cummins, B., & Laga, I. (2025) · Social Networks · doi:10.1016/j.socnet.2025.07.005 · Cited by 1
Harvest restrictions fail to influence population abundance
Gurney, S. M., Christensen, S. A., Nichols, M. J., Stewart, C. M., Williams, D. M., Mayhew, S. L., Gilbert, N. A., & Etter, D. R. (2025) · Ecosphere · doi:10.1002/ecs2.70358 · Cited by 1
Itter, M. S., & Finley, A. O. (2025) · Ecological Applications · doi:10.1002/eap.70084 · Cited by 5
“Among species standard deviation parameters ( τ ℓ ) are assigned half-Cauchy prior distributions (Gelman, 2006).”
The Gut Check: On Stability and Homogeneity Methods for Reference Materials
Gierz, K. (2025) · Journal of Statistical Theory and Practice · doi:10.1007/s42519-025-00490-y · Cited by 2
“This distribution is weakly informative and has been suggested as an alternative to a half-Cauchy [43, 44].”
Bayesian Optimal Phase II (BOP2) Design for Basket Trials
He, T., Ren, Y., & Zhou, H. (2025) · Statistics in Biopharmaceutical Research · doi:10.1080/19466315.2025.2546355 · Cited by 1
Skew Gaussian Markov Random Fields Under Decomposable Graphs
Zareifard, H., & Khaledi, M. J. (2025) · Environmetrics · doi:10.1002/env.70039 · Cited by 2
“Gelman (2006) noted the sensitivity of inference to the choice of 𝑎 and suggested a weakly informative uniform prior density on the standard deviation parameter.”
An early-life survival and reproductive trade-off shapes selection on body size
Janeiro, M. J., Henshaw, J. M., Pemberton, J. M., et al. (2025) · Evolution Letters · doi:10.1093/evlett/qraf029 · Cited by 1
Nguyen, P. H., Engel, S. M., & Herring, A. H. (2025) · International Journal of Environmental Research and Public Health · doi:10.3390/ijerph22101466 · Cited by 2
“…by the latent factors of X i , the variation due to unobserved factors and idiosyncratic noise: where ν 2 has a half-t prior with a small degree of freedom for an uninformative prior that still behaves well in the case that between-subject variance ν 2 is close to zero, as suggested by [32].”
Experimental observation of the prior residence advantage in masu salmon in an experimental channel
Yamashita, Y., Iwasaki, Y., & Strüssmann, C. A. (2025) · Ecological Solutions and Evidence · doi:10.1002/2688-8319.70132 · Cited by 1
Meta-analysis with a single study
van Zwet, E. W., Więcek, W., & Gelman, A. (2025) · Statistical Methods in Medical Research · doi:10.1177/09622802251380628 · Cited by 1
Self-citationModel-Quality-Adaptive Probability of Detection Across Multiple Inspection Scenarios
Scheirer, N., & Holland, S. D. (2025) · Journal of Nondestructive Evaluation · doi:10.1007/s10921-025-01286-7 · Cited by 1
“The half-Cauchy was selected to avoid the numerical instabilities from a noninfor-mative lognormal distribution in a hierarchical model [47].”
Focal thalamic infrared neural stimulation propagates dynamical transformations in auditory cortex
Coventry, B. S., Luu, C. P., & Bartlett, E. L. (2025) · Journal of Neural Engineering · doi:10.1088/1741-2552/ae1bdb · Cited by 1
A multilevel model with heterogeneous variances for snap timing in the National Football League
Nguyen, Q., & Yurko, R. (2025) · Journal of the Royal Statistical Society Series A (Statistics in Society) · doi:10.1093/jrsssa/qnaf185 · Cited by 3
“We use the following weakly informative prior distributions for the standard deviation parameters in our model: where half-t 3 represents a half-t distribution (i.e., the absolute value of a Student-t distribution centered at zero) with 3 degrees of freedom (see Gelman (2006) for more details).”
Curtis, J. S., Dillingham, P., & Wing, S. R. (2025) · Scientific Reports · doi:10.1038/s41598-025-23130-8 · Cited by 1
Heath, A., Rios, D., Vogel, K. I., Rowe, T., Wills-Ibarra, N., Oakley, E. A., Offringa, M., Pechlivanoglou, P., Klassen, T. P., Dalziel, S. R., et al. (2025) · Trials · doi:10.1186/s13063-025-09125-w · Cited by 1
Cost-to-coast: space–time modeling of coast defense funds and environmental drivers in Italy
Ricciotti, L., Pollice, A., Picone, M., & Martino, S. (2025) · Environmental and Ecological Statistics · doi:10.1007/s10651-025-00686-2 · Cited by 1
“The choice of priors in Bayesian modeling is a topic of active research, particularly for hierarchical models (Gelman 2006; de Zea Bermudez et al. 2024 ).”
Shiny-MAGEC: A Bayesian R shiny application for meta-analysis of censored adverse events
Zhou, Z., Tian, Z., Peterson, C. B., Bao, L., & Zhou, S. (2025) · Research Synthesis Methods · doi:10.1017/rsm.2025.10052 · Cited by 1
“Such weakly-informative prior with its scale parameter A equal to 2.5, 10, or 25 was generally recommended for the SD parameters in hierarchical models.”
InfluentialTrinder, S. A., Hidalgo, O., Fay, M. F., et al. (2025) · Ecology and Evolution · doi:10.1002/ece3.72438 · Cited by 1
“The prior for variance components was a non-informative uniform improper prior distribution on the standard deviation of the random effects (specified as V = 1.0 × 10 − 16 , nu = − 1 in mcmcglmm), as recommended by Gelman (2006).”
InfluentialMulder, J., & van Aert, R. C. M. (2025) · Research Synthesis Methods · doi:10.1017/rsm.2025.10060 · Cited by 6
“(cid:5) vi Under the marema model, the Berger–Deely prior would be 𝑝 79,83 should be used with care, as they may unduly be highly noninformative.”
Monteiro, A. M., Schneider, A., Encarnação, S., Forte, P. M., Barbosa, T. M., & Pecos‐Martín, D. (2025) · Systematic Reviews · doi:10.1186/s13643-025-02899-9 · Cited by 2
“The Half-Cauchy distribution was chosen as a prior for between-study variance in the Bayesian framework [32, 33], as it provides a weakly informative yet robust prior, particularly suitable for modeling uncertainty in variance parameters.”
Seemingly unrelated Bayesian additive regression trees for cost-effectiveness analyses in healthcare
Eßer, J., Maia, M., Parnell, A. C., Bosmans, J. E., van Dongen, J. M., Klausch, T. L. T., & Murphy, K. (2025) · The Annals of Applied Statistics · doi:10.1214/25-aoas2087 · Cited by 2
“See Chip-man et al. (2010), Kapelner and Bleich (2016), and Tan and Roy (2019) for further details of the model and additional information the algorithmic implementation on which that of the suBART models is based.”
A Bayesian physics-informed approach to inverse problems in heat transfer and fluid mechanics
Fernandez, R. C., Zanini, C. T. P., Schmidt, A. M., et al. (2025) · Applied Mathematical Modelling · doi:10.1016/j.apm.2025.116638 · Cited by 1
Weusten, J. J. A. M., & Kim, J. Y. (2025) · Pharmaceutical Research · doi:10.1007/s11095-025-03979-0 · Cited by 1
Evolving spatiotemporal patterns and urban scaling of deaths from external causes
Filho, C. I. N. S., Carmona, H. A., Neto, A. S. L., Prates, M. V., Ribeiro, H. V., Castro, M. C., & Andrade, J. S. (2026) · Scientific Reports · doi:10.1038/s41598-025-31970-7 · Cited by 2
Investigating Momentum Effects in Home Prices: Evidence from Transaction-Level Data in South Korea*
Pyo, D.-J. (2025) · Housing Finance Research · doi:10.52344/hfr.2025.9.1.59 · Cited by 1
Drivers of population dynamics and juvenile mortality in Northwest Atlantic harp seals
Tinker, M. T., Stenson, G. B., Mosnier, A., de Walle, J. V., Lang, S. L. C., & Hammill, M. O. (2026) · Ecological Applications · doi:10.1002/eap.70184 · Cited by 2
“Methods for the LSA were based on those described by Eacker et al. (2017) and are described in detail in Appendix S1: Section S1.”
Bayesian Estimation of Contagion Effect: An Application of Friendship Networks and Alcohol Behavior
Ndreka, B., Dey, D. K., & Lachos, V. H. (2026) · The New England Journal of Statistics in Data Science · doi:10.51387/26-nejsds94 · Cited by 1
Consistent and Scalable Variable Selection with Robust Link Functions
Odoom, E., & Wang, X. (2026) · The New England Journal of Statistics in Data Science · doi:10.51387/26-nejsds102 · Cited by 1
Carmo, M., Williams, D. R., & Rast, P. (2026) · Journal of Educational and Behavioral Statistics · doi:10.3102/10769986261426004 · Cited by 1
Covariate Adjustment in Basket Trials Borrowing Information Across Subgroups
Ren, J., Robertson, D. S., & Zheng, H. (2026) · Statistics in Medicine · doi:10.1002/sim.70492 · Cited by 1
“We refer the interested reader to Gelman [47] and Cunanan et al. [46] for general recommendations on priors for variance parameters.”
Su, W., Wang, X., Miranda-Afonso, P., Andrinopoulou, E.-R., & Szczesniak, R. D. (2026) · The New England Journal of Statistics in Data Science · doi:10.51387/26-nejsds99 · Cited by 1
Parker, J., Williams, C. L., Deel, H. L., Letitiya, D., Goldenberg, S. Z., Sukuna, D. A., D’Ammando, G., Daballen, D., Owen, M. A., Tubbs, C. W., et al. (2026) · Royal Society Open Science · doi:10.1098/rsos.251569 · Cited by 1
Design and analysis of experiments assessing handheld RFID readers in complex environments
Strait, J., Weaver, B., Dorn, M. F., et al. (2026) · Journal of Quality Technology · doi:10.1080/00224065.2026.2639400 · Cited by 1
Sathasivam, M., Allen, D. K., Shankar, V., Saha, R., & Narayanan, S. (2026) · Journal of Experimental Botany · doi:10.1093/jxb/erag183 · Cited by 5
Tolerance and prediction intervals: The Bayesian way
Ramírez, J. G., & Chen, F. (2026) · Journal of Biopharmaceutical Statistics · doi:10.1080/10543406.2026.2670527 · Cited by 2
“Common types include: (1) Noninformative (Diffuse) priors, where the distribution is “flat” relative to the likelihood function; used when little prior knowledge exists (Gelman 2006).”
Liu, G., Liang, Z., Wang, W., & Ge, R. (2026) · Frontiers in Physiology · doi:10.3389/fphys.2026.1777058 · Cited by 1
“Speci fi cally, weakly informative or non-informative normal (0, s = 1) priors were assigned to the fi xed effects (Gelman, 2006; Gelman et al., 2013).”
Disentangling species-specific controls on urban tree transpiration: A Bayesian framework
Zhang, X., Zhong, X., Yan, J., Li, Q., Yao, L.-Y., Liu, K., Moser-Reischl, A., Rötzer, T., Pauleit, S., & Rahman, M. A. (2026) · Agricultural and Forest Meteorology · doi:10.1016/j.agrformet.2026.111381 · Cited by 1
Time-Course Microarray Data Analysis
Xu, S. (2012) · Springer eBooks · doi:10.1007/978-0-387-70807-2_23 · Cited by 2
Xu, S. (2012) · Springer eBooks · doi:10.1007/978-0-387-70807-2_6 · Cited by 1
A Bayesian hierarchical model for forecasting intermountain snow dynamics
Odei, J. B., Symanzik, J., & Hooten, M. B. (2014) · Environmetrics · doi:10.1002/env.2275 · Cited by 2
“With these types of models, variances are notoriously sensitive to prior specifications (Gelman, 2006). Thus, how well the model behaves would depend on good choices of initial values for the variances. In this article, we chose values that were relatively diffused (spread out) and conservative for what we thought the…”
BackgroundInfluentialA Multilevel Regression Model for Geographical Studies in Sets of Non-Adjacent Cities
Marí-Dell’Olmo, M., & Martínez‐Beneito, M. A. (2015) · PLoS ONE · doi:10.1371/journal.pone.0133649 · Cited by 1
Najafabadi, A. T. P. (2015) · Cogent Mathematics · doi:10.1080/23311835.2015.1023670 · Cited by 1
“Gelman (2006) constructed a non-central t student family of conditionally conjugate priors for hierarchical standard deviation parameters.”
MethodInfluentialWu, S., Black, M. A., North, R. A., & Rodrigo, A. G. (2012) · BMC Bioinformatics · doi:10.1186/1471-2105-13-137 · Cited by 1
van den Berg, S. M. (2008) · Behavior Genetics · doi:10.1007/s10519-008-9239-7 · Cited by 1
“The prior distribution for the square root of the phenotypic variance was uniformly distributed (Gelman, 2006) and constrained to lie between 0.5 and 1.5.”
BackgroundInfluentialA single latent trait model for multiple binary outcomes in a cluster randomized clinical trial
Zhao, X., Stone, R. A., Ye, F., & Fine, M. J. (2011) · Health Services and Outcomes Research Methodology · doi:10.1007/s10742-011-0078-2 · Cited by 1
“For the variance parameters of the random effects, we used noninformative uniform priors on standard deviation (SD) parameters, as recommended by Gelman (2006). Uniform (0,100) priors were used for r, s, and {wh}, the SDs of random effects at the patient, provider and site levels, respectively. We assume that all of t…”
MethodInfluentialXu, S. (2012) · Springer eBooks · doi:10.1007/978-0-387-70807-2_15 · Cited by 2
Predator–prey models: an application for the plankton dynamics of lake Geneva
Achcar, J. A., Mazucheli, J., & Coelho-Barros, E. A. (2010) · Environmental and Ecological Statistics · doi:10.1007/s10651-010-0134-z · Cited by 1
“In this case, we assumed uniform prior distributions for the variance components (Gelman 2006), but the obtained DIC values were larger when compared to the DIC values assuming Gamma prior distributions.”
Toti, S., Biggeri, A., & Baldasseroni, A. (2006) · Statistical Modelling · doi:10.1191/1471082x06st110oa · Cited by 1
A zero-inflated Bayesian spatial model with near-locating measurements
Zhang, J., & He, C. Z. (2013) · Environmental and Ecological Statistics · doi:10.1007/s10651-013-0237-4 · Cited by 1
“And Gelman (2006) suggests using a uniform prior or the half-t family on the hierarchical standard deviation, rather than the inversegamma (ε, ε) family with a very small ε.”
MethodTopics on Dynamic Panel Data Models with Random Effects Using Semi-Parametric Bayesian Approach
Kazemi, I., & Rikhtehgaran, R. (2014) · Communication in Statistics- Theory and Methods · doi:10.1080/03610926.2012.673672 · Cited by 1
“…and the prior of σ 2α is assumed uniform distribution with the range 0–100 which is similar to an inverse-Gamma prior with hyper-parameters 0.01 and 0.01 (Gelman, 2006); for the parameter ϕ a normal prior with zero mean and precision 0.001 is assumed; for regression coefficients, we assume each…”
BackgroundInfluentialLi, G., Richardson, S. T., Fortunato, L., et al. (2011) · IEEE ... International Conference on Data Mining workshops · doi:10.1109/icdmw.2011.59 · Cited by 1
“In this section, we recall the construction of BaySTDetect [7] in the context of disease surveillance, although the detection framework can be adapted to other types of small area data such as crime-mapping.”
MethodComment on Article by Rubio and Steel
Scott, J. G. (2014) · Bayesian Analysis · doi:10.1214/13-ba867 · Cited by 1
Model Comparison, Model Checking, and Hypothesis Testing
Cowles, M. K. (2012) · Springer texts in statistics · doi:10.1007/978-1-4614-5696-4_11 · Cited by 1
Evaluation of multilevel decision trees
Rogard, E., Gelman, A., & Lü, H. (2006) · Journal of Statistical Planning and Inference · doi:10.1016/j.jspi.2006.01.011 · Cited by 1
Self-citationZhang, Y., Lowe, D. P., Briggs, D., et al. (2015) · IFAC-PapersOnLine · doi:10.1016/j.ifacol.2015.10.147 · Cited by 1
Unified variable selection in semi-parametric models
Terry, W., Zhang, H., Maity, A., Arshad, S. H., & Karmaus, W. (2015) · Statistical Methods in Medical Research · doi:10.1177/0962280215610928 · Cited by 1
“The fourth prior distribution is a uniform distribution from 0 to 100, Uniform(0, 100), a non-informative but proper prior suggested in Gelman (2006).20 We apply each of the four prior distributions to the 500 MC replicates and record the three statistics (percentage of correct selection,…”
MethodA BAYESIAN APPROACH TO JOINT MODELING OF MENSTRUAL CYCLE LENGTH AND FECUNDITY
Lum, K. J., Sundaram, R., Louis, G. M. B., Louis, T. A., Sundaram, K. J., Buck-louis, R., Germaine, M., Louis, T. A., Joint, A. B. A. T., & Louis, T. A. (2014) · Collection of Biostatistics Research Archive · Cited by 1
A Comparison of Bayesian Models of Heteroscedasticity in Nested Normal Data
Herschtal, A., Foroudi, F., Kron, T., & Mengersen, K. L. (2015) · Communications in Statistics - Simulation and Computation · doi:10.1080/03610918.2014.936467 · Cited by 1
“Both the real world datasets and the artificial datasets are described in Section 2.6 Gelman (2006) also considers candidates for prior distributions for variance parameters in hierarchical normal models, however our work differs from his in two important ways.”
InfluentialVelasco‐Cruz, C., Contreras-Cruz, L. F., Smith, E. P., & Rodríguez, J. E. (2016) · Journal of Agricultural Biological and Environmental Statistics · doi:10.1007/s13253-016-0250-9 · Cited by 1
Kruschke, J. K. (2014) · Elsevier eBooks · doi:10.1016/b978-0-12-405888-0.00024-6 · Cited by 1
A Bayesian Model to Assess $$T_2$$ Values and Their Changes Over Time in Quantitative MRI
Combès, B., Kerbrat, A., Commowick, O., et al. (2016) · Lecture notes in computer science · doi:10.1007/978-3-319-46726-9_66 · Cited by 1
Multivariate hierarchical Bayesian models and choice of priors in the analysis of survey data
Montaño, D. (2016) · Journal of Applied Statistics · doi:10.1080/02664763.2016.1267120 · Cited by 1
Ranking ultimate teams using a Bayesian score-augmented win-loss model
Murray, T. A. (2017) · Journal of Quantitative Analysis in Sports · doi:10.1515/jqas-2016-0097 · Cited by 1
BAYESIAN ANALYSIS IN A SPEEDED ITEM RESPONSE MODEL
Bazán, J. L., Valdivieso, L., & Branco, M. D. (2010) · Cited by 1
Metadata
- Authors
- Gelman, A.
- Affiliations
- Columbia University (US)
- Corresponding author
- Gelman, Andrew
- ORCID
- Gelman, Andrew
- Venue
- Bayesian Analysis
- Publisher
- Institute of Mathematical Statistics
- Published
- 2006-09-01
- Volume
- 1
- Issue
- 3
- Pages
- 515-534
- ISSN
- 1931-6690 (ISSN-L)
- 1936-0975
- OpenAlex
- W1536497620
- Semantic Scholar
- 63d40805d4635b70f6b5aa07fc2c9821480cfc1e
- Cited by
- 4,050
- References
- 34
- Citations by year
- 2026: 194 · 2025: 218 · 2024: 237 · 2023: 237 · 2022: 282 · 2021: 312 · 2020: 320 · 2019: 281 · 2018: 266 · 2017: 263 · 2016: 235 · 2015: 217 · 2014: 176 · 2013: 142 · 2012: 184
- FWCI
- 78.70
- Type
- journal article
- Language
- en
- Open access
- diamond
- doi.org
- Licence
- cc-by
- Locations
- Bayesian Analysis (publishedVersion, cc-by)
- Project Euclid (Cornell University) (submittedVersion)
- Funders
- National Science Foundation
- Topics
- Physical Sciences › Mathematics › Statistics and Probability › Statistical Methods and Bayesian Inference
- Physical Sciences › Mathematics › Statistics and Probability › Statistical Distribution Estimation and Applications
- Physical Sciences › Computer Science › Artificial Intelligence › Bayesian Methods and Mixture Models
- Keywords
- prior distributions, hierarchical models, scale parameter, non-informative priors, weakly informative priors, uniform prior, analysis of variance
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-10-02, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-02 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.