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 2,301–2,400 of 4,089
Godinho, D. P., Fragata, I., Majer, A., Rodrigues, L. R., & Magalhães, S. (2024) · Journal of Evolutionary Biology · doi:10.1093/jeb/voae003 · Cited by 6
The Flexible Gumbel Distribution: A New Model for Inference about the Mode
Liu, Q., Huang, X., & Zhou, H. (2024) · Stats · doi:10.3390/stats7010019 · Cited by 9
“We choose the above prior for the scale parameters by following the prior selection for variance parameters suggested in Gelman [2006].”
MethodModeling the Intraindividual Relation of Ability and Speed within a Test
Mutak, A., Krause, R. W., Ulitzsch, E., Much, S., Ranger, J., & Pohl, S. (2024) · Journal of Educational Measurement · doi:10.1111/jedm.12391 · Cited by 7
Catalog variance of testing general relativity with gravitational-wave data
Pacilio, C., Gerosa, D., & Bhagwat, S. (2024) · Physical review. D/Physical review. D. · doi:10.1103/physrevd.109.l081302 · Cited by 6
Bayesian Scalable Precision Factor Analysis for Gaussian Graphical Models
Chandra, N. K., Müller, P., & Sarkar, A. (2024) · Bayesian Analysis · doi:10.1214/24-ba1461 · Cited by 18
Inferring Covariance Structure from Multiple Data Sources via Subspace Factor Analysis
Chandra, N. K., Dunson, D. B., & Xu, J. (2024) · Journal of the American Statistical Association · doi:10.1080/01621459.2024.2408777 · Cited by 22
Inverse Gaussian mixtures models of bearing vibrations under local faults
Boškoski, P., & Juričić, Đ. (2015) · Mechanical Systems and Signal Processing · doi:10.1016/j.ymssp.2015.05.010 · Cited by 6
A Bayesian-Hierarchical Space-Time Model for Significant Wave Height Data
Vanem, E., Huseby, A. B., & Natvig, B. (2011) · doi:10.1115/omae2011-49716 · Cited by 6
Lawson, A. B., Onicescu, G., & Meinzer, C. (2011) · Spatial and Spatio-temporal Epidemiology · doi:10.1016/j.sste.2011.07.004 · Cited by 6
Torabi, M. (2013) · Journal of Applied Statistics · doi:10.1080/02664763.2013.847066 · Cited by 6
Kallache, M., Maksimovich, E., Michelangeli, P., & Naveau, P. (2010) · Journal of Climate · doi:10.1175/2010jcli3107.1 · Cited by 5
“A uniform prior for the precision l is proportional to a uniform prior for the standard deviation s (see, e.g., Gelman 2006).”
MethodA Bayesian Semiparametric Multiplicative Error Model With an Application to Realized Volatility
Solgi, R., & Mira, A. (2012) · Journal of Computational and Graphical Statistics · doi:10.1080/10618600.2013.810151 · Cited by 6
“This strategy (sample from the posterior of the PX-model using an MCMC simulation, and then post-process the path of the Markov chain to obtain a sample from the posterior of the original model) was also used in Gelman (2006).”
MethodA Latent Class Model with Hidden Markov Dependence for Array CGH Data
DeSantis, S. M., Houseman, E. A., Coull, B. A., et al. (2009) · Biometrics · doi:10.1111/j.1541-0420.2009.01226.x · Cited by 6
“Thus, rather than a relatively noninformative gamma prior for this parameter, a more informative prior is a U(0,1), i.e., one that allows for double the observed variance (Gelman, 2006).”
BackgroundChen, H., Bakshi, B. R., & Goel, P. K. (2009) · AIChE Journal · doi:10.1002/aic.11918 · Cited by 6
Leroux, A. J., & Beretvas, S. N. (2018) · Methodology · doi:10.1027/1614-2241/a000143 · Cited by 6
Model Comparison and Simulation for Hierarchical Models: Analyzing Rural–Urban Migration in Thailand
Garip, F., & Western, B. (2011) · Chapman & Hall/CRC eBooks · doi:10.1201/b10905-25 · Cited by 6
Powell, H., & Lee, D. (2013) · Journal of the Royal Statistical Society Series A (Statistics in Society) · doi:10.1111/rssa.12034 · Cited by 6
“Finally, the spatial variance parameter σ2 is typically assigned a conjugate inverse gamma prior distribution, but this has been shown to be informative for small values of σ2 (Gelman, 2006).”
BackgroundGreen, C., Schmitz, J. M., Lindsay, J. A., Pedroza, C., Lane, S. D., Agnelli, R., Kjome, K., & Moeller, F. G. (2012) · Frontiers in Psychiatry · doi:10.3389/fpsyt.2012.00092 · Cited by 7
“There has been some concern that such priors may not be appropriate (see Gelman, 2006 for a discussion in the context of normal distributed data).”
BackgroundApproximate uniform shrinkage prior for a multivariate generalized linear mixed model
Chen, H.-C., & Wehrly, T. E. (2015) · Journal of Multivariate Analysis · doi:10.1016/j.jmva.2015.12.004 · Cited by 6
Yu, Y., & Wang, X. (2015) · Statistics in Medicine · doi:10.1002/sim.6830 · Cited by 6
Variable selection in semi-parametric models
Zhang, H., Maity, A., Arshad, S. H., Holloway, J. W., & Karmaus, W. (2013) · Statistical Methods in Medical Research · doi:10.1177/0962280213499679 · Cited by 6
Bayesian spatially dependent variable selection for small area health modeling
Choi, J., & Lawson, A. B. (2016) · Statistical Methods in Medical Research · doi:10.1177/0962280215627184 · Cited by 6
Jenatabadi, H. S., Babashamsi, P., Khajeheian, D., & Seyyedamiri, N. (2016) · Sustainability · doi:10.3390/su8111204 · Cited by 6
“Weakly informative priors can be recommended when the researchers want to apply a weakerer prior than what your actual knowledge would allow [62].”
BackgroundEstimating abundance without recaptures of marked pallid sturgeon in the Mississippi River
Friedenberg, N. A., Hoover, J. J., Boysen, K. A., & Killgore, K. J. (2017) · Conservation Biology · doi:10.1111/cobi.12972 · Cited by 6
Kim, K.-H., Cho, Y.-S., & Kwon, H. (2017) · Stochastic Environmental Research and Risk Assessment · doi:10.1007/s00477-017-1488-7 · Cited by 6
“Hence, non-informative prior distributions for the regression parameters are assigned, as suggested in the literatures (Gelman 2006; Gelman et al. 2014).”
MethodBayesian Models for Weighted Data with Missing Values: A Bootstrap Approach
Goldstein, H., Carpenter, J. R., & Kenward, M. G. (2018) · Journal of the Royal Statistical Society Series C (Applied Statistics) · doi:10.1111/rssc.12259 · Cited by 6
“We note, however, that, in the context of multilevel models where the number of higher level clusters is not large, then the choice of default priors for the corresponding variance parameters needs careful consideration (Gelman, 2006).”
BackgroundYazdani, A., Yazdani, A., Yazdani, A., et al. (2019) · Scientific Reports · doi:10.1038/s41598-019-41362-3 · Cited by 6
“Therefore, in large scale problems, this choice of prior yields to highly concentrated distribution about the scale matrix(4)(5)(6).”
BackgroundAreal prediction of survey data using Bayesian spatial generalised linear models
Bakar, K. S., & Jin, H. (2018) · Communications in Statistics - Simulation and Computation · doi:10.1080/03610918.2018.1530787 · Cited by 6
Naranjo, L., Pérez, C. J., Fuentes–García, R., et al. (2019) · Biostatistics · doi:10.1093/biostatistics/kxz004 · Cited by 6
AdaptSPEC-X: Covariate-Dependent Spectral Modeling of Multiple Nonstationary Time Series
Bertolacci, M., Rosen, O., Cripps, E., & Cripps, S. (2021) · Journal of Computational and Graphical Statistics · doi:10.1080/10618600.2021.2000870 · Cited by 8
Hepp, G., & Zessner, M. (2019) · Water · doi:10.3390/w11102161 · Cited by 6
Kiang, M. V., Krieger, N., Buckee, C. O., Onnela, J.-P. “., & Chen, J. T. (2019) · BMJ Open · doi:10.1136/bmjopen-2019-029373 · Cited by 8
“6 7 Following Gelman, 24 we reparameterised our precisions as SD with priors of σ = 1 .”
Bayesian Deconvolution and Quantification of Metabolites from J-Resolved NMR Spectroscopy
Heinecke, A., Ye, L., Iorio, M. D., & Ebbels, T. M. D. (2020) · Bayesian Analysis · doi:10.1214/20-ba1208 · Cited by 7
“As suggested by Gelman (2006), we also assign a half Cauchy distribution to the global shrinkage parameter, τ ∼ C+(0, d).”
MethodKaluskar, S., Blukacz‐Richards, E. A., Johnson, C. A., Kim, D., & Arhonditsis, G. B. (2020) · Ecological Complexity · doi:10.1016/j.ecocom.2020.100846 · Cited by 7
Sampling Strategies for Fast Updating of Gaussian Markov Random Fields
Brown, D. A., Watson Self, S., & McMahan, C. S. (2019) · TigerPrints (Clemson University) · Cited by 6
Chen, Y., Lu, Y., & Moustaki, I. (2022) · The Annals of Applied Statistics · doi:10.1214/21-aoas1564 · Cited by 10
Conditional adaptive Bayesian spectral analysis of replicated multivariate time series
Li, Z., Bruce, S. A., Wutzke, C. J., et al. (2021) · Statistics in Medicine · doi:10.1002/sim.8884 · Cited by 6
Bayesian hierarchical evaluation of dose-response for peanut allergy in clinical trial screening
Haber, L. T., Reichard, J. F., Henning, A. K., et al. (2021) · Food and Chemical Toxicology · doi:10.1016/j.fct.2021.112125 · Cited by 6
“Following Gelman et al. (2006), we have used half-Cauchy distributions for the standard deviation hyper-parameters (sigmai and sigmac) in all models.”
MethodJay, M., Oleson, J., Charlton, M. E., & Arab, A. (2021) · Statistics in Medicine · doi:10.1002/sim.8948 · Cited by 8
A warming Southern Ocean may compromise Antarctic blue whale foetus growth
Smith, C. (2021) · Folia Zoologica · doi:10.25225/jvb.20114 · Cited by 7
Bayesian hierarchical modeling of operating room times for surgeries with few or no historic data
Gañan-Cardenas, E., Jiménez, J. C., & Pemberthy-R., J. I. (2021) · Journal of Clinical Monitoring and Computing · doi:10.1007/s10877-021-00696-y · Cited by 6
“Regarding the modeling of the scale parameters related to the random effects, we evaluate the uniform model, the half-Cauchy model [14], and the ScaleBeta2 model [21].”
MethodInfluentialFast approximate inference for multivariate longitudinal data
Hughes, D. M., García‐Fiñana, M., & WAND, M. P. (2021) · Biostatistics · doi:10.1093/biostatistics/kxab021 · Cited by 6
A heterogeneous Bayesian regression model for skewed spatial data
Zareifard, H., & Khaledi, . J. (2021) · Spatial Statistics · doi:10.1016/j.spasta.2021.100545 · Cited by 6
Salmerón, D., Botta, L., Martínez, J. M., Trama, A., Gatta, G., Borràs, J. M., Capocaccia, R., Clèries, R., & for the Information Network on Rare Cancers (RARECARENet) Working Group (2021) · American Journal of Epidemiology · doi:10.1093/aje/kwab262 · Cited by 7
“We compared the performance of 13 computational approaches (see Web Appendices 3–5) for estimating the mean number of cases and their confidence/credible intervals by using maximum likelihood approaches, empirical Bayes approaches, and modeling (12–22).”
Rawson, T. (2022) · Food Microbiology · doi:10.1016/j.fm.2021.103975 · Cited by 7
Joo, S., & Lee, P. (2022) · Journal of Educational Measurement · doi:10.1111/jedm.12316 · Cited by 6
A meta-analytic assessment of a Thyroglobulin marker for marbling in beef cattle
Wood, I., Moser, G., Burrell, D. L., et al. (2006) · Genetics Selection Evolution · doi:10.1051/gse:2006016 · Cited by 6
Creus-Martí, I., Moyá, A., & Santonja, F. J. (2022) · Complexity · doi:10.1155/2022/4907527 · Cited by 6
“As [55] suggests, σ ij is assigned the uniform distribution in the interval and then E ( y it ) can be estimated by E ( y it ) � α it / � 1 α it As we have mentioned before, the numerator and the denominator of the principal balances are selected so as to maximize the variance of the balance.”
König, C., Spoden, C., & Frey, A. (2022) · Behavior Research Methods · doi:10.3758/s13428-022-02000-5 · Cited by 6
Kümmerli, R., & Keller, L. (2011) · The American Naturalist · doi:10.1086/659951 · Cited by 6
“V p 10^2, which represents the prior standard deviation with a scale of 10 (Gelman 2006; Hadfield 2010).”
BackgroundThe use of asymmetric distributions in average bioequivalence
de Souza, R. M., Achcar, J. A., Martínez, E. Z., & Mazucheli, J. (2016) · Statistics in Medicine · doi:10.1002/sim.6885 · Cited by 6
“In presence of random effects, we consider a hierarchical Bayesian analysis [39].”
MethodDimensional and statistical foundations for accumulated damage models
Wong, S. W. K., & Zidek, J. V. (2017) · Wood Science and Technology · doi:10.1007/s00226-017-0953-2 · Cited by 6
“These choices of priors represent the absence of any prior knowledge on the parameters (Gelman 2006).”
BackgroundMultidimensional multiple group IRT models with skew normal latent trait distributions
Gomez, J. L. P., Azevedo, C., & Lachos, V. H. (2018) · Journal of Multivariate Analysis · doi:10.1016/j.jmva.2018.05.003 · Cited by 7
A Bayesian hierarchical model for analyzing methylated RNA immunoprecipitation sequencing data
Zhang, M., Li, Q., & Xie, Y. (2018) · Quantitative Biology · doi:10.1007/s40484-018-0149-2 · Cited by 7
“One standard way of setting a weakly informative gamma prior is to choose small values for the two hyperparameters, such as af=bf=0:001 [20].”
BackgroundSu, Y., Feng, Q., Zhu, G., Gu, C., Wang, Y., Shang, S., Zhang, K., Han, T., Chen, H., & Ma, J. (2018) · Hydrological Processes · doi:10.1002/hyp.13298 · Cited by 6
“For the residual variance of measurement error in Equation (13), because there is enough data available to estimate them, we can just assume independent uniform priors (Gelman, 2006), p σ2s ∝ 1.”
BackgroundA Bayesian Estimation of Child Labour in India
Kim, J., Olsen, W. K., & Wiśniowski, A. (2020) · Child Indicators Research · doi:10.1007/s12187-020-09740-w · Cited by 6
“Gelman (2006) suggests three different priors: uniform, inverse gamma, and half Cauchy.”
BackgroundModelling the impact of MAUP on environmental drivers for Schistosoma japonicum prevalence
Navas, A. A., Osei, F., Magalhães, R. J. S., Leonardo, L., & Stein, A. (2020) · Parasites & Vectors · doi:10.1186/s13071-020-3987-5 · Cited by 7
Nonparametric Bayesian inference for the spectral density based on irregularly spaced data
Zhang, S. (2020) · Computational Statistics & Data Analysis · doi:10.1016/j.csda.2020.107019 · Cited by 6
A BAYESIAN JOINT MODEL FOR POPULATION AND PORTFOLIO-SPECIFIC MORTALITY
van Berkum, F., Antonio, K., & Vellekoop, M. (2017) · Astin Bulletin · doi:10.1017/asb.2017.17 · Cited by 11
Horrocks, G., Cunningham, S. C., O’Dowd, D. J., Thomson, J. R., & Nally, R. M. (2011) · Austral Ecology · doi:10.1111/j.1442-9993.2011.02315.x · Cited by 7
A sign based loss approach to model selection in nonparametric regression
Nott, D. J., & Li, J. (2009) · Statistics and Computing · doi:10.1007/s11222-009-9139-6 · Cited by 6
“See Gelman (2006) for some further discussion of priors on variance parameters in complex hierarchical models.”
BackgroundStepBrothers: inferring partially shared ancestries among recombinant viral sequences
Bloomquist, E. W., Dorman, K. S., & Suchard, M. A. (2008) · Biostatistics · doi:10.1093/biostatistics/kxn019 · Cited by 8
“The distributions for (κ j , µ j ) are independent of each other given the hyperparameters φ (Minin and others, 2005; Gelman, 2006).”
BackgroundCabrini, S. M., Irwin, S. H., & Good, D. (2010) · American Journal of Agricultural Economics · doi:10.1093/ajae/aaq024 · Cited by 6
“shrinkage intensity. Combining the sample likelihood derived from equation ( 7 ) with the prior”
Kim, J., & Park, J. (2016) · Ocean & Coastal Management · doi:10.1016/j.ocecoaman.2016.10.014 · Cited by 7
A Unit Level Small Area Model with Misclassified Covariates
Arima, S., & Polettini, S. (2019) · Journal of the Royal Statistical Society Series A (Statistics in Society) · doi:10.1111/rssa.12468 · Cited by 6
“Gelman (2006) pointed out that inverse gamma priors for the scale parameters in hierarchical models cannot be considered as non-informative.”
BackgroundTipton, J., Hooten, M. B., Nolan, C. J., et al. (2019) · The Annals of Applied Statistics · doi:10.1214/19-aoas1281 · Cited by 6
“mendations in Gelman (2006) and assign independent half-Cauchy priors on the standard deviations τ , where for j = 1, .”
MethodBayesian Survival Analysis of Type I General Exponential Distributions
AbuJarad, M. H. A., AbuJarad, E. S. A., & Khan, A. A. (2019) · Annals of Data Science · doi:10.1007/s40745-019-00228-1 · Cited by 6
Disease Mapping Using Bayesian Hierarchical Models
Earnest, A., Cramb, S. M., & White, N. M. (2012) · Wiley series in probability and statistics · doi:10.1002/9781118394472.ch13 · Cited by 6
Inefficiency of Data Augmentation for Large Sample Imbalanced Data.Preprint
Johndrow, J. E., Smith, A. D., Pillai, N. S., & Dunson, D. B. (2016) · arXiv (Cornell University) · Cited by 8
Varying-Censoring Aware Matrix Factorization for Single Cell RNA-SequencingPreprint
Townes, F. W., Hicks, S. C., Aryee, M. J., & Irizarry, R. A. (2017) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/166736 · Cited by 6
de Miguel, M., Rodríguez‐Quilón, I., Heuertz, M., et al. (2020) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2020.03.02.974113 · Cited by 6
A Bayesian Nonlinear Mixed-Effects Location Scale Model for LearningPreprint
Williams, D. R., & Rast, P. (2018) · doi:10.31219/osf.io/r438f · Cited by 6
Prior sensitivity analysis in default Bayesian structural equation modelingPreprint
van Erp, S., Mulder, J., & Oberski, D. L. (2017) · PsyArXiv (OSF Preprints) · doi:10.31234/osf.io/5j3m9 · Cited by 7
Bayesian hierarchical modeling of size spectraPreprint
Wesner, J. S., Pomeranz, J. P. F., Junker, J. R., et al. (2023) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2023.02.14.528491 · Cited by 6
“…or time, the simple model above can be expanded to include predictors and/or varying intercepts and slopes: The literature on prior choice is broad and active (Banner et al. 2020; Wesner & Pomeranz 2021), particularly for priors on hyperparameters like 𝜎 𝑗 (Gelman 2006; Aguilar & Bürkner 2023).”
BackgroundWang, Z., Murray, T. A., Xiao, M., Lin, L., Alemayehu, D., & Chu, H. (2023) · Statistics in Medicine · doi:10.1002/sim.9902 · Cited by 6
“We do not recommend non-informative priors since they may result in slow convergence of the MCMC sampling algorithm, and the diffuse normal priors for the fixed-effect parameters on the logit-transformed probability parameters may actually provide strong information on the probability scale.”
BackgroundD’Orangeville, L., Itter, M. S., Santos, J. M. D., & Taylor, A. R. (2024) · Forest Ecology and Management · doi:10.1016/j.foreco.2024.122482 · Cited by 6
Statistical flaws of the fitness-fatigue sports performance prediction model
Marchal, A., Benazieb, O., Weldegebriel, Y., Méline, T., & Imbach, F. (2025) · Scientific Reports · doi:10.1038/s41598-025-88153-7 · Cited by 10
Integrative cardiovascular dose–response to graded lower‐body negative pressure
Whittle, R. S., Keller, N., Hall, E. A., Patanam, S., Dunbar, B. J., & Diaz‐Artiles, A. (2025) · Experimental Physiology · doi:10.1113/ep092483 · Cited by 7
“Following the recommendations of Gelman (2006), half-Cauchy distributionswereusedforall 𝜎 and 𝜎 u .”
Wang, Q., Xu, P., Zhang, K., et al. (2025) · Transportmetrica A Transport Science · doi:10.1080/23249935.2025.2511816 · Cited by 5
Ye, Y., Zheng, P., Xu, P., Ren, Q., Yan, R., & Gao, X. (2025) · Reliability Engineering & System Safety · doi:10.1016/j.ress.2025.111690 · Cited by 5
Huang, J., Li, Y., Brellenthin, A. G., Lee, D.-C., Sui, X., & Blair, S. N. (2021) · Journal of Applied Statistics · doi:10.1080/02664763.2021.1962260 · Cited by 5
“A LKJ-Correlation prior [27] is assigned to R1 and half-Student-t priors [13] are assigned to variance parameters in σ 1.”
MethodA Bayesian hierarchical model for individual participant data meta‐analysis of demand curves
Zhang, S., Chu, H., Bickel, W. K., Le, C. T., Smith, T. T., Thomas, J., Donny, E. C., Hatsukami, D. K., & Luo, X. (2022) · Statistics in Medicine · doi:10.1002/sim.9354 · Cited by 5
Bayesian sparse modeling to identify high‐risk subgroups in meta‐analysis of safety data
Qi, X., Zhou, S., Wang, Y., & Peterson, C. B. (2022) · Research Synthesis Methods · doi:10.1002/jrsm.1597 · Cited by 6
Bayesian borrowing for basket trials with longitudinal outcomes
Whitehead, L. E., Sailer, O., Witham, M. D., & Wason, J. M. S. (2023) · Statistics in Medicine · doi:10.1002/sim.9751 · Cited by 7
“A prior distribution, g k , is placed on the shrinkage parameter, which is set to a function of the HD, d kk ∗ : As discussed in Reference 22, g k is chosen as a spike and slab distribution (the main reason being that it allows for strong borrowing given a basket with sufficiently consistent treatment effect, and weak…”
MethodInfluentialZambrano, H., Torres, X., Coleman, M., Franchi, M. V., Fisher, J. P., Oberlin, D. J., Hooren, B. V., Swinton, P. A., & Schoenfeld, B. J. (2023) · Scientific Reports · doi:10.1038/s41598-023-35424-w · Cited by 5
Jaya, I. G. N. M., Handoko, B., Andriyana, Y., Chadidjah, A., Kristiani, F., & Antikasari, M. (2023) · Mathematics · doi:10.3390/math11173641 · Cited by 7
Ardakani, O. M. (2024) · Econometrics and Statistics · doi:10.1016/j.ecosta.2024.04.002 · Cited by 5
A hierarchical signal detection model with unequal variance for binary responses
Lages, M. (2024) · Psychonomic Bulletin & Review · doi:10.3758/s13423-024-02504-5 · Cited by 5
Ultra-Efficient MCMC for Bayesian Longitudinal Functional Data Analysis
Sun, T. Y., & Kowal, D. R. (2024) · Journal of Computational and Graphical Statistics · doi:10.1080/10618600.2024.2362227 · Cited by 5
Meta-analysis of Censored Adverse Events
Qi, X., Zhou, S., Peterson, C. B., Wang, Y., Fang, X., Wang, M. L., & Shen, C. (2024) · The New England Journal of Statistics in Data Science · doi:10.51387/24-nejsds62 · Cited by 10
Design and analysis of cluster randomized trials
Li, W., Xie, Y., Pham, D., Dong, N., Spybrook, J. K., & Kelcey, B. M. (2024) · Asia Pacific Education Review · doi:10.1007/s12564-024-09984-z · Cited by 11
“Other studies provided alternative methods (e.g., generalized estimating equations, cluster bootstrapping, Bayesian methods, etc.) to analyze CRTs with a small number of clusters (e.g., Gelman, 2006; Huang, 2016; Morel et al., 2003).”
Tomita, F. (2024) · Earth Planets and Space · doi:10.1186/s40623-024-02120-6 · Cited by 5
Debusho, L. K., & Gemechu, L. L. (2025) · BMC Public Health · doi:10.1186/s12889-024-20996-7 · Cited by 5
Atlas, W. I., Glaser, D. M., Connors, B. M., Holt, C. A., Greenberg, D. A., Selbie, D. T., Cox‐Rogers, S., Carr‐Harris, C., Hertz, E., Hart, J. L., et al. (2025) · Marine and Coastal Fisheries · doi:10.1093/mcfafs/vtaf008 · Cited by 5
Large reaching datasets quantify the impact of age, sex/gender, and experience on motor control
Zhang, A., Ruitenberg, M. F. L., Warburton, M., Scott, S., & Tsay, J. S. (2026) · Communications Psychology · doi:10.1038/s44271-025-00383-7 · Cited by 3
“These priors are chosentobe weakly informative so that the data drive the inference with reasonable constraints 40 – 42 .”
Estimating brood‐specific reproductive inhibition potency in aquatic toxicity testing
Zhang, J., Bailer, A. J., & Oris, J. T. (2012) · Environmetrics · doi:10.1002/env.2181 · Cited by 5
“Following the suggestion of Gelman (2006), rather than using inverse gamma prior distributions (Inverse Gamma.”
MethodEstimation Performance for the Bayesian Hierarchical Linear Model
Korso, M. N. E., Boyer, R., Larzabal, P., et al. (2016) · IEEE Signal Processing Letters · doi:10.1109/lsp.2016.2528579 · Cited by 5
Hierarchical Bayes small‐area estimation with an unknown link function
Sugasawa, S., Kubokawa, T., & Rao, J. N. K. (2018) · Scandinavian Journal of Statistics · doi:10.1111/sjos.12376 · Cited by 5
“For the two scale parameters A and 𝜆, using an inverse gamma distribution is a typical choice, but the use of an inverse gamma distribution might be sensitive in hierarchical models (Gelman, 2006).”
MethodYielding stability of early maturing potato varieties: Bayesian analysis
Przystalski, M., & Lenartowicz, T. (2020) · The Journal of Agricultural Science · doi:10.1017/s0021859620000945 · Cited by 6
“Instead, it was proposed to use half-Cauchy distribution or inverse-gamma distribution as a prior for variance components (Gelman, 2006) and scaled inverse Wishart distribution (Gelman et al., 2014) as prior for covariance matrices (Gelman et al., 2014).”
MethodLeiva-Yamaguchi, V., & Alvares, D. (2020) · Entropy · doi:10.3390/e23010050 · Cited by 5
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.