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 401–500 of 4,089
Kokubu, S., Nishihara, G. N., Watanabe, Y., Tsuchiya, Y., Amamo, Y., & Terada, R. (2015) · Phycologia · doi:10.2216/15-007.1 · Cited by 67
“Weakly informative normal priors were placed on all of the parameters of the model, and a half-cauchy prior was placed on the scale parameter of the models (Gelman 2004, 2006).”
MethodA Flexible Hierarchical Bayesian Modeling Technique for Risk Analysis of Major Accidents
Yu, H., Khan, F. I., & Veitch, B. (2017) · Risk Analysis · doi:10.1111/risa.12736 · Cited by 79
“The noninformative prior can be updated using data D to form an informative posterior of the parameter of interest.(23) This informative posterior is derived through three basic steps.”
MethodAhn, W., Haines, N., & Zhang, L. (2016) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/064287 · Cited by 67
Ambati, J., Magagnoli, J., Leung, H., Wang, S.-B., Andrews, C. A., Fu, D., Pandey, A., Sahu, S., Narendran, S., Hirahara, S., et al. (2020) · Nature Communications · doi:10.1038/s41467-020-18528-z · Cited by 66
Stepwise-edited, human melanoma models reveal mutations’ effect on tumor and microenvironment
Hodis, E., Triglia, E. T., Kwon, J., Biancalani, T., Zakka, L. R., Parkar, S., Hütter, J.-C., Buffoni, L., Delorey, T. M., Phillips, D., et al. (2022) · Science · doi:10.1126/science.abi8175 · Cited by 70
Zeng, Q., Wang, Q., Zhang, K., Wong, S. C., & Xu, P. (2023) · Accident Analysis & Prevention · doi:10.1016/j.aap.2023.107119 · Cited by 73
A new role for effort dynamics in the theory of harvested populations and data-poor stock assessment
Thorson, J. T., Minto, C., Minte‐Vera, C. V., Kleisner, K. M., & Longo, C. (2013) · Canadian Journal of Fisheries and Aquatic Sciences · doi:10.1139/cjfas-2013-0280 · Cited by 70
Probabilistic numerical methods for PDE-constrained Bayesian inverse problems
Cockayne, J., Oates, C. J., Sullivan, T. J., et al. (2017) · AIP conference proceedings · doi:10.1063/1.4985359 · Cited by 65
“The length-scale hyper-parameter ` was incorporated into the MCMC procedure, endowed with a half-Cauchy hyper-prior as recommended in [6].”
MethodSome models for stochastic frontiers with endogeneity
Griffiths, W. E., & Hajargasht, G. (2015) · Journal of Econometrics · doi:10.1016/j.jeconom.2015.06.012 · Cited by 66
“A gamma prior for 2 is in line with most traditional priors specified for variance parameters, while use of a uniform prior for standard deviations in hierarchical models (which bear some similarity to our model) has been advocated by Gelman (2006).”
MethodGoldsmith, J., & Kitago, T. (2015) · Journal of the Royal Statistical Society Series C (Applied Statistics) · doi:10.1111/rssc.12115 · Cited by 78
“Although these priors are convenient for the development of a straightforward Gibbs sampler and for the derivation of the variational approximation in Section 2.2, they have been criticized by several researchers (Gelman, 2006; Yang and Berger, 1994).”
InfluentialGuinet, C., Xing, X., Walker, E. A., Monestiez, P., Marchand, S., Picard, B., Jaud, T., Authier, M., Cotté, C., Dragon, A., et al. (2013) · Earth system science data · doi:10.5194/essd-5-15-2013 · Cited by 73
High-throughput functional evaluation of BRCA2 variants of unknown significance
Ikegami, M., Kohsaka, S., Ueno, T., Momozawa, Y., Inoue, S., Tamura, K., Shimomura, A., Hosoya, N., Kobayashi, H., Tanaka, S., et al. (2020) · Nature Communications · doi:10.1038/s41467-020-16141-8 · Cited by 68
An informed reference prior for between‐study heterogeneity in meta‐analyses of binary outcomes
Pullenayegum, E. M. (2011) · Statistics in Medicine · doi:10.1002/sim.4326 · Cited by 64
Xu, X., Xie, S., Wong, S. C., Xu, P., Huang, H., & Pei, X. (2016) · Journal of Advanced Transportation · doi:10.1002/atr.1442 · Cited by 68
Prior specification in Bayesian occupancy modelling improves analysis of species occurrence data
Outhwaite, C. L., Chandler, R. E., Powney, G. D., Collen, B., Gregory, R. D., & Isaac, N. J. B. (2018) · Ecological Indicators · doi:10.1016/j.ecolind.2018.05.010 · Cited by 79
“The half-Cauchy is the same as the Student’s t distribution with 1 degree of freedom but restricted to positive values, and it has been recommended for use as the default prior for scale parameters within the literature (Gelman, 2006; Polson and Scott, 2012).”
MethodImpact of Tanzania’s Wildlife Management Areas on household wealth
Keane, A. M., Lund, J. F., Bluwstein, J., Burgess, N. D., Nielsen, M. R., & Homewood, K. (2019) · Nature Sustainability · doi:10.1038/s41893-019-0458-0 · Cited by 80
Distance to native climatic niche margins explains establishment success of alien mammals
Broennimann, O., Petitpierre, B., Chevalier, M., González‐Suárez, M., Jeschke, J. M., Rolland, J., Gray, S. M., Bacher, S., & Guisan, A. (2021) · Nature Communications · doi:10.1038/s41467-021-22693-0 · Cited by 78
The tempo of cetacean cranial evolution
Coombs, E. J., Felice, R. N., Clavel, J., Park, T., Bennion, R. F., Churchill, M., Geisler, J. H., Beatty, B. L., & Goswami, A. (2022) · Current Biology · doi:10.1016/j.cub.2022.04.060 · Cited by 74
A Bayesian multinomial model to analyse spatial patterns of childhood co-morbidity in Malawi
Kazembe, L. N., & Namangale, J. J. (2007) · European Journal of Epidemiology · doi:10.1007/s10654-007-9145-y · Cited by 63
The genetic covariance between life cycle stages separated by metamorphosis
Aguirre, J. D., Blows, M. W., & Marshall, D. J. (2014) · Proceedings of the Royal Society B Biological Sciences · doi:10.1098/rspb.2014.1091 · Cited by 66
Bayesian Factor Analysis as a Variable-Selection Problem: Alternative Priors and Consequences
Lu, Z., Chow, S., & Loken, E. (2016) · Multivariate Behavioral Research · doi:10.1080/00273171.2016.1168279 · Cited by 71
“This hierarchical normal mixture setting is widely used in Bayesian variable selection and shrinkage estimates (Gelman, 2006).”
MethodInvited Article: Concepts and tools for the evaluation of measurement uncertainty
Possolo, A. M., & Iyer, H. K. (2017) · Review of Scientific Instruments · doi:10.1063/1.4974274 · Cited by 72
Mensink, P. J., & King, K. (2019) · British Journal of Educational Technology · doi:10.1111/bjet.12752 · Cited by 76
Cadonna, A., Frühwirth‐Schnatter, S., & Knaus, P. (2020) · Econometrics · doi:10.3390/econometrics8020020 · Cited by 63
Shut and re-open: the role of schools in the spread of COVID-19 in Europe
Stage, H. B., Shingleton, J., Ghosh, S., Scarabel, F., Pellis, L., & Finnie, T. (2021) · Philosophical Transactions of the Royal Society B Biological Sciences · doi:10.1098/rstb.2020.0277 · Cited by 66
Estimation methods for nonlinear state-space models in ecology
Pedersen, M. W., Berg, C. W., Thygesen, U. H., Nielsen, A., & Madsen, H. (2011) · Ecological Modelling · doi:10.1016/j.ecolmodel.2011.01.007 · Cited by 62
AdaptSPEC: Adaptive Spectral Estimation for Nonstationary Time Series
Rosen, O., Wood, S. A., & Stoffer, D. S. (2012) · Journal of the American Statistical Association · doi:10.1080/01621459.2012.716340 · Cited by 62
“To complete the prior specification on g ( ν ), we follow Gelman (2006) and assume τ 2 ∼ U (0 , c τ 2 ), where c τ 2 is a known large value.”
MethodHierarchical shrinkage priors for dynamic regressions with many predictors
Korobilis, D. (2012) · International Journal of Forecasting · doi:10.1016/j.ijforecast.2012.05.006 · Cited by 78
Scale-Dependent Priors for Variance Parameters in Structured Additive Distributional Regression
Klein, N., & Kneib, T. (2015) · Bayesian Analysis · doi:10.1214/15-ba983 · Cited by 77
Ishikura, K., Yamada, H., Toma, Y., Takakai, F., Morishita, T., Darung, U., Limin, A., Limin, S. H., & Hatano, R. (2016) · Soil Science & Plant Nutrition · doi:10.1080/00380768.2016.1244652 · Cited by 63
“HalfCauchy distribution (Gelman 2006) was applied for γαj, γβj, δαk, δβk, λαl, λβl, σy, σα and σβ, and normal distribution was applied for µα and µβ.”
Fitch, D. T., Sharpnack, J., & Handy, S. (2020) · Transportation Research Part F Traffic Psychology and Behaviour · doi:10.1016/j.trf.2020.02.015 · Cited by 66
Does online food delivery improve the equity of food accessibility? A case study of Nanjing, China
Zhang, S., Luan, H., Zhen, F., Kong, Y., & Xi, G. (2023) · Journal of Transport Geography · doi:10.1016/j.jtrangeo.2022.103516 · Cited by 76
Climate-driven variation in dispersal ability predicts responses to forest fragmentation in birds
Weeks, T. L., Betts, M. G., Pfeifer, M., Wolf, C., Banks-Leite, C., Barbaro, L., Barlow, J., Cerezo, A., Kennedy, C. M., Kormann, U. G., et al. (2023) · Nature Ecology & Evolution · doi:10.1038/s41559-023-02077-x · Cited by 72
Semiparametric Bayes hierarchical models with mean and variance constraints
Yang, M., Dunson, D. B., & Baird, D. D. (2010) · Computational Statistics & Data Analysis · doi:10.1016/j.csda.2010.03.025 · Cited by 56
“Gelman (2006) induces a prior on the variance of a random effect in a parametric model by expressing the random effect as a transformation of a latent variable in an over-parameterized, or parameter-expanded (PE), model.”
BackgroundMethodInfluentialApproximate inference for disease mapping with sparse Gaussian processes
Vanhatalo, J. P., Pietiläinen, V., & Vehtari, A. (2010) · Statistics in Medicine · doi:10.1002/sim.3895 · Cited by 74
Hierarchical Bayesian modelling of rail track geometry degradation
Andrade, A. R., & Teixeira, P. F. (2013) · Proceedings of the Institution of Mechanical Engineers Part F Journal of Rail and Rapid Transit · doi:10.1177/0954409713486619 · Cited by 67
MODELING SPATIAL PATTERNS IN FISHERIES BYCATCH: IMPROVING BYCATCH MAPS TO AID FISHERIES MANAGEMENT
Sims, M., Cox, T. M., & Lewison, R. L. (2008) · Ecological Applications · doi:10.1890/07-0685.1 · Cited by 63
Phillis, C. C., Moore, J. W., Buoro, M., Hayes, S. A., Garza, J. C., & Pearse, D. E. (2015) · Journal of Heredity · doi:10.1093/jhered/esv085 · Cited by 63
“…inverse-Wishart matrix distribution with 2 degrees of freedom as a prior for the residual variance; for the variance components we use a parameter expanded prior composed of an inverse-gamma distribution with 2 degrees of freedom and a normal distribution with mean 0 and variance 625 (Gelman 2006).”
MethodInfluentialLiu, F., & Eugenio, E. C. (2016) · Statistical Methods in Medical Research · doi:10.1177/0962280216650699 · Cited by 83
Effects of the North Pacific Current on the productivity of 163 Pacific salmon stocks
Malick, M. J., Cox, S., Mueter, F., Dorner, B., & Peterman, R. (2017) · Fisheries Oceanography · doi:10.1111/fog.12190 · Cited by 48
Fluke, J., Harlaar, N., Brown, B., Heisler, K. W., Merkel‐Holguin, L., & Darnell, A. J. (2018) · Child Maltreatment · doi:10.1177/1077559518816381 · Cited by 63
Objectifying Contextual Effects. The Use of Snowball Sampling in Political Sociology
Audemard, J. (2020) · Bulletin of Sociological Methodology/Bulletin de Méthodologie Sociologique · doi:10.1177/0759106319888703 · Cited by 101
Zhang, X., Gu, X., Cheng, C., & Yang, D. (2020) · The Science of The Total Environment · doi:10.1016/j.scitotenv.2020.140925 · Cited by 57
Genetic variation in recombination rate in the pig
Johnsson, M., Whalen, A., Ros‐Freixedes, R., Gorjanc, G., Chen, C.-Y., Herring, W. O., de Koning, D., & Hickey, J. M. (2021) · Genetics Selection Evolution · doi:10.1186/s12711-021-00643-0 · Cited by 61
“We used parameter expanded priors [49] for the variance of permanent environmental effects and for the additive genetic effects, using V = 1, ν = 1, αμ = 0, αV = 1000, which corresponds to a half-Cauchy prior with a scale of 100, and an inverse-Wishart prior (V = 1, ν = 1) for the residual variance.”
MethodThorlund, K., Thabane, L., & Mills, E. J. (2013) · BMC Medical Research Methodology · doi:10.1186/1471-2288-13-2 · Cited by 60
Watanabe, Y., Nishihara, G. N., Tokunaga, S., & Terada, R. (2014) · Phycological Research · doi:10.1111/pre.12053 · Cited by 66
“Weakly informative normal priors were placed on all of the parameters of the model, and a half-cauchy prior was placed on the scale parameter of the models (Gelman 2004, 2006).”
MethodAlexeeff, S., Baccarelli, A., Halonen, J. I., Coull, B. A., Wright, R. O., Tarantini, L., Bollati, V., Sparrow, D., Vokonas, P., & Schwartz, J. (2013) · International Journal of Epidemiology · doi:10.1093/ije/dys220 · Cited by 62
Efthimiou, O., Mavridis, D., Cipriani, A., Leucht, S., Bagos, P., & Salanti, G. (2014) · Statistics in Medicine · doi:10.1002/sim.6117 · Cited by 60
The paradox of inverted biomass pyramids in kelp forest fish communities
Trebilco, R., Dulvy, N. K., Anderson, S. C., & Salomon, A. K. (2016) · Proceedings of the Royal Society B Biological Sciences · doi:10.1098/rspb.2016.0816 · Cited by 63
Growth rate, transmission mode and virulence in human pathogens
Leggett, H. C., Cornwallis, C. K., Buckling, A., & West, S. A. (2017) · Philosophical Transactions of the Royal Society B Biological Sciences · doi:10.1098/rstb.2016.0094 · Cited by 66
Lima, C., Kwon, H., & Kim, Y. (2018) · Journal of Hydrology · doi:10.1016/j.jhydrol.2018.08.075 · Cited by 61
Chevalier, M., Russell, J. C., & Knape, J. (2018) · Ecological Applications · doi:10.1002/eap.1838 · Cited by 72
Liu, M., Bunn, V., Hupf, B., Lin, J., & Lin, J. (2021) · Statistics in Medicine · doi:10.1002/sim.9095 · Cited by 75
Preferred policy responses to technological change: Survey evidence from OECD countries
Busemeyer, M. R., Gandenberger, M., Knotz, C., & Tober, T. (2022) · Socio-Economic Review · doi:10.1093/ser/mwac015 · Cited by 83
Bayesian Estimation of Population-Level Trends in Measures of Health Status
Finucane, M. M., Paciorek, C. J., Danaei, G., & Ezzati, M. (2014) · Statistical Science · doi:10.1214/13-sts427 · Cited by 59
“We give the precision parameters a flat prior on the standard deviation scale (Gelman, 2006), truncating logλ ≤ 15, as larger values correspond to essentially no extra-linear temporal variability.”
MethodInfluentialBayesian approaches in evolutionary quantitative genetics
O’Hara, R. B., Cano, J. M., Ovaskainen, O., Teplitsky, C., & ALHO, J. S. (2008) · Journal of Evolutionary Biology · doi:10.1111/j.1420-9101.2008.01529.x · Cited by 58
“Recent investigations, following on from Gelman (2006), suggest that especial care should be taken with choosing an uninformative prior, but that noninformative and weakly informative priors for variance components (such as additive genetic variance) with good frequentist properties are available…”
BackgroundInfluentialBridger, D., Bonner, S. J., & Briffa, M. (2015) · Proceedings of the Royal Society B Biological Sciences · doi:10.1098/rspb.2014.2492 · Cited by 62
Iversen, E. S., Couch, F. J., Goldgar, D. E., Tavtigian, S. V., & Monteiro, Á. N. A. (2011) · Cancer Epidemiology Biomarkers & Prevention · doi:10.1158/1055-9965.epi-10-1214 · Cited by 64
On the Use of Co-Data in Clinical Trials
Neuenschwander, B. E., Roychoudhury, S., & Schmidli, H. (2016) · Statistics in Biopharmaceutical Research · doi:10.1080/19466315.2016.1174149 · Cited by 62
“This follows advice by Spiegelhalter and Gelman (Spiegelhalter, Abrams, and Myles 2004; Gelman 2006), who proposed half-normal, half-Cauchy, and half-t distributions with appropriate scale parameters.”
MethodHespanhol, L., van Mechelen, W., & Verhagen, E. A. L. M. (2017) · British Journal of Sports Medicine · doi:10.1136/bjsports-2016-097025 · Cited by 63
Seasonal Food Scarcity Prompts Long-Distance Foraging by a Wild Social Bee
Pope, N. S., & Jha, S. (2017) · The American Naturalist · doi:10.1086/694843 · Cited by 68
Preliminary examination of the orexin system on relapse-related factors in cocaine use disorder
Suchting, R., Yoon, J. H., Miguel, G. G. S., Green, C. E., Weaver, M. F., Vincent, J. N., Fries, G. R., Schmitz, J. M., & Lane, S. D. (2019) · Brain Research · doi:10.1016/j.brainres.2019.146359 · Cited by 61
Alternating imputation posterior estimation of models with crossed random effects
Cho, S., & Rabe‐Hesketh, S. (2010) · Computational Statistics & Data Analysis · doi:10.1016/j.csda.2010.04.015 · Cited by 59
“Gelman (2006) suggested using the half-t or half-Cauchy distribution in this case.”
The Court of Justice and treaty revision: A case of strategic leniency?
Castro-Montero, J. L., Alblas, E., Dyèvre, A., et al. (2018) · European Union Politics · doi:10.1177/1465116518793707 · Cited by 58
“Following the recommendations of Gelman (2006) and Bürkner (2016), we choose conservative priors for coefficients and intercepts.”
Personality and Prosocial Behavior: A Multilevel Meta-Analysis
Kline, R., Bankert, A., Levitan, L. C., & Kraft, P. (2017) · Political Science Research and Methods · doi:10.1017/psrm.2017.14 · Cited by 59
Lima, C., & Lall, U. (2009) · Water Resources Research · doi:10.1029/2008wr007485 · Cited by 58
A Bayesian approach to identifying and compensating for model misspecification in population models
Thorson, J. T., Ono, K., & Munch, S. B. (2013) · Ecology · doi:10.1890/13-0187.1 · Cited by 59
Fraterrigo, J. M., Wagner, S., & Warren, R. J. (2014) · Ecology Letters · doi:10.1111/ele.12352 · Cited by 63
Liu, H., Zhang, Z. J., & Grimm, K. J. (2015) · Structural Equation Modeling A Multidisciplinary Journal · doi:10.1080/10705511.2015.1057285 · Cited by 67
“Gelman (2006) used τ = 25 for weak-informative Cauchy distribution.”
BackgroundMethodInfluentialEffects of intranasal oxytocin on symptoms of schizophrenia: A multivariate Bayesian meta-analysis
Williams, D. R., & Bürkner, P. (2016) · Psychoneuroendocrinology · doi:10.1016/j.psyneuen.2016.10.013 · Cited by 68
Fast and flexible linear mixed models for genome-wide genetics
Runcie, D. E., & Crawford, L. A. (2019) · PLoS Genetics · doi:10.1371/journal.pgen.1007978 · Cited by 70
“However, others have recommended uniform or half-t-family priors for 319 the standard-deviation parameters of hierarchical models [62], which are easily 320 implemented in Stan [52].”
BackgroundChiu, W. A., Lynch, M. T., Lay, C. R., Antezana, A., Malek, P., Sokolinski, S., & Rogers, R. D. (2022) · Environmental Health Perspectives · doi:10.1289/ehp10103 · Cited by 57
“The prior distribution for population variation was based on a weakly informative half-normal prior (per a recommendation by Gelman 2006) with mean of 0.16.”
InfluentialWho Wants What?: Redistribution Preferences in Comparative Perspective
Rueda, D., & Stegmueller, D. (2019) · Cited by 59
Lack of ownership of mobile phones could hinder the rollout of mHealth interventions in Africa
Okano, J. T., Ponce, J., Krönke, M., & Blower, S. M. (2022) · eLife · doi:10.7554/elife.79615 · Cited by 62
“We used weakly informative priors (Gelman, 2006; Gelman et al., 2013; Gelman et al., 2008; Lemoine, 2019; Williams et al., 2018):”
MethodInfluentialPaciorek, C. J., & McLACHLAN, J. S. (2009) · Journal of the American Statistical Association · doi:10.1198/jasa.2009.0026 · Cited by 66
“In particular, for variance components, we have used uniform priors on the standard deviation scale to avoid the use of diffuse inverse gamma priors (Gelman 2006), which have sharp spikes in density at small values, and decay extremely rapidly to zero density at values smaller than the location of…”
Authier, M., Bentaleb, I., Ponchon, A., Martin, C., & Guinet, C. (2012) · PLoS ONE · doi:10.1371/journal.pone.0032026 · Cited by 48
“Weakly informative priors were used [94,95].”
Bayesian forecasting of parts demand
Yelland, P. M. (2010) · International Journal of Forecasting · doi:10.1016/j.ijforecast.2009.11.001 · Cited by 66
“Finally, we should point out that in order to achieve reasonable fidelity to the data, we have adopted a so-called hierarchical or multilevel prior (Gelman 2006; Gelman and Hill 2006) in the model.”
MethodAppropriate vegetation indices for measuring the impacts of deer on forest ecosystems
Iijima, H., & Nagaike, T. (2014) · Ecological Indicators · doi:10.1016/j.ecolind.2014.09.009 · Cited by 61
Spatial heterogeneity in the carrying capacity of sika deer in Japan
Iijima, H., & Ueno, M. (2016) · Journal of Mammalogy · doi:10.1093/jmammal/gyw001 · Cited by 60
Retrieval-based model accounts for striking profile of episodic memory and generalization
Banino, A., Koster, R., Hassabis, D., & Kumaran, D. (2016) · Scientific Reports · doi:10.1038/srep31330 · Cited by 57
A Bayesian Approach for Estimating Multilevel Latent Contextual Models
Zitzmann, S., Lüdtke, O., Robitzsch, A., & Marsh, H. W. (2016) · Structural Equation Modeling A Multidisciplinary Journal · doi:10.1080/10705511.2016.1207179 · Cited by 68
“…it is often less critical to select priors for regression coefficients, there is an ongoing discussion about which prior distribution should be specified for variance components (e.g., Browne & Draper, 2006; Chung, Rabe-Hesketh, Dorie, Gelman, & Liu, 2013; Gelman, 2006; Natarajan & Kass, 2000).”
BackgroundBest, N., & Hansell, A. (2009) · Epidemiology · doi:10.1097/ede.0b013e31819d90f9 · Cited by 55
Yi, N., Liu, N., Zhi, D., & Li, J. (2011) · PLoS Genetics · doi:10.1371/journal.pgen.1002382 · Cited by 62
A comparative analysis of Chikungunya and Zika transmission
Riou, J., Poletto, C., & Boëlle, P. (2017) · Epidemics · doi:10.1016/j.epidem.2017.01.001 · Cited by 59
The spatial econometrics of the coronavirus pandemic
Krisztin, T., Piribauer, P., & Wögerer, M. (2020) · Letters in Spatial and Resource Sciences · doi:10.1007/s12076-020-00254-1 · Cited by 63
“5 It is worth noting that that the standard choice of an inverted gamma prior for the error nuisance might arguably severely affect the results (see, for example, Gelman 2006; Simpson et al.”
BackgroundEstimation of the Biological Half-Life of Methylmercury Using a Population Toxicokinetic Model
Jo, S., Woo, H. D., Kwon, H., Oh, S., Park, J., Hong, Y., Pyo, H., Park, K., Ha, M., Kim, H., et al. (2015) · International Journal of Environmental Research and Public Health · doi:10.3390/ijerph120809054 · Cited by 56
“and set a uniform prior between 0 and 100 for σ as a non-informative prior [25].”
Bayesian model selection: The steepest mountain to climb
Tenan, S., O’Hara, R. B., Hendriks, I. E., & Tavecchia, G. (2014) · Ecological Modelling · doi:10.1016/j.ecolmodel.2014.03.017 · Cited by 58
Effects of the Fukushima Daiichi nuclear accident on goshawk reproduction
Murase, K., Murase, J., Horie, R., & Endo, K. (2015) · Scientific Reports · doi:10.1038/srep09405 · Cited by 61
Munoz, F., Violle, C., & Cheptou, P. (2015) · Oikos · doi:10.1111/oik.02328 · Cited by 61
“We further used parameter expansion to improve MCMC performance (Gelman 2006, Hadfield 2014).”
MethodXie, S., Dong, N., Wong, S. C., Huang, H., & Xu, P. (2018) · Accident Analysis & Prevention · doi:10.1016/j.aap.2018.09.030 · Cited by 61
“In accordance with Gelman (2006), the variance parameters, i.e., u , V , P , V , and P were assigned as a uniform (0,10) .”
Sampling-based inference of the primordial CMB and gravitational lensing
Millea, M., Anderes, E., & Wandelt, B. D. (2020) · Physical review. D/Physical review. D. · doi:10.1103/physrevd.102.123542 · Cited by 61
Tustison, N. J., Holbrook, A. J., Avants, B., Roberts, J. M., Cook, P. A., Reagh, Z. M., Duda, J., Stone, J. R., Gillen, D. L., Yassa, M. A., et al. (2019) · Journal of Alzheimer s Disease · doi:10.3233/jad-190283 · Cited by 86
Learning Hamiltonian Monte Carlo in R
Thomas, S., & Tu, W. (2020) · The American Statistician · doi:10.1080/00031305.2020.1865198 · Cited by 79
“R> plot(fm1_hmc, burnin=200) Note that the Inverse Gamma distribution is not always an optimal prior when the support is strictly positive.”
MethodInfluentialMulti-Modal Analgesic Strategy for Trauma: A Pragmatic Randomized Clinical Trial
Harvin, J. A., Albarado, R. P., Truong, V. T. T., Green, C., Tyson, J. E., Pedroza, C., Wade, C. E., Kao, L. S., Hudson, J. A., Taub, E. A., et al. (2021) · Journal of the American College of Surgeons · doi:10.1016/j.jamcollsurg.2020.12.014 · Cited by 57
Reporting Practice in Multilevel Modeling: A Revisit After 10 Years
Luo, W., Li, H., Baek, E., Chen, S., Lam, K. H., & Semma, B. (2021) · Review of Educational Research · doi:10.3102/0034654321991229 · Cited by 76
“Bayesian estimation computes the joint posterior distributions of the fixed-effect parameter and the variance components of the random effects (Gelman, 2006; Zeger & Karim, 1991) based on the prior distributions and the data.”
BackgroundMethodBois, F. Y. (2012) · Methods in molecular biology · doi:10.1007/978-1-62703-059-5_25 · Cited by 69
Comparison of REML and Gibbs sampling estimates of multi-trait genetic parameters in Scots pine
Waldmann, P., & Ericsson, T. (2006) · Theoretical and Applied Genetics · doi:10.1007/s00122-006-0246-x · Cited by 53
“Moreover, it has recently been argued that the inverse gamma family (which includes the inverse chi-square and inverse Wishart distributions) can be problematic as prior for variances and result in biased estimates (Gelman 2006).”
BackgroundAir pollution and health in Scotland: a multicity study
Lee, D., Ferguson, C., & Mitchell, R. (2009) · Biostatistics · doi:10.1093/biostatistics/kxp010 · Cited by 60
“The default variance prior for λ adopted in this paper is uniform on the standard deviation scale as suggested by Gelman (2006), rather than the more commonly used inverse gamma( , ) distribution for = 0.001, 0.01, 0.1.”
InfluentialFirmat, C., Delzon, S., Louvet, J., Parmentier, J., & Kremer, A. (2017) · Journal of Evolutionary Biology · doi:10.1111/jeb.13185 · Cited by 55
“The contribution of microevolution to short-term responses to new climatic conditions remains a matter of debate (Gienapp et al., 2008; Meril€a, 2012; Franks et al., 2014; Teplitsky & Millien, 2014).”
Uncertainty-driven regulation of learning and exploration in adolescents: A computational account
Jepma, M., Schaaf, J. V., Visser, I., & Huizenga, H. M. (2020) · PLOS Computational Biology · doi:10.1371/journal.pcbi.1008276 · Cited by 53
“These priors and hyperpriors were chosen for their 982 uninformative nature [64].”
Background
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.