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 3,001–3,100 of 4,089
Data-Driven Prior Distributions for A Bayesian Phase-2 COPD Dose-Finding Clinical Trial
Novick, S. J., Ho, S., & Best, N. (2018) · Statistics in Biopharmaceutical Research · doi:10.1080/19466315.2018.1462728 · Cited by 2
A Bayesian approach to errors-in-variables beta regression
Figueroa-Zúñiga, J., Carrasco, J. M. F., Arellano-Valle, R. B., et al. (2018) · Brazilian Journal of Probability and Statistics · doi:10.1214/17-bjps354 · Cited by 2
“As suggested by Gelman (2006) the prior distribution φ = U2 with U ∼ U(0, a) with large a (a = 50 for example) is less informative than an inverse gamma prior; the prior distribution for φ we suggest is then a more flexible version of Gelman’s prior as shown in Fig.”
MethodMetric Predicted Variable with One Nominal Predictor
Kruschke, J. K. (2014) · Elsevier eBooks · doi:10.1016/b978-0-12-405888-0.00019-2 · Cited by 2
Interval Estimations for Variance Components: A Review and Implementations
Quiroz, J., & Baumgartner, R. (2019) · Statistics in Biopharmaceutical Research · doi:10.1080/19466315.2018.1512896 · Cited by 2
“Another prior information for σ 2A proposed by Gelman (2006) is the half-Cauchy family prior with location parameter of 0 and scale parameter of 25 .”
MethodA hybrid model for simulating grazing herds in real time
Demšar, J., Blewitt, W., & Bajec, I. L. (2019) · Computer Animation and Virtual Worlds · doi:10.1002/cav.1914 · Cited by 3
Muchmore, P., Xu, S., Marjoram, P., Rappaport, E. B., Weng, J., Molshatzki, N., & Eckel, S. P. (2020) · Physiological Reports · doi:10.14814/phy2.14336 · Cited by 1
“Overestimating the standard deviation results in posterior distributions that are over dispersed, or equivalently leads to credible intervals that are wider than they would otherwise be (Gelman, 2006).”
BackgroundCastaño-Díez, C., Castaño, L. S. Á., Caicedo-Velásquez, B., et al. (2020) · Revista chilena de nutrición · doi:10.4067/s0717-75182020000100022 · Cited by 2
Estimation of group means using Bayesian generalized linear mixed models
Lalonde, A., & Qu, Y. (2020) · Pharmaceutical Statistics · doi:10.1002/pst.2006 · Cited by 1
Equity of access to critical care services in Scotland: A Bayesian spatial analysis
Emerson, P., Green, D. R., Stott, S., MacLennan, G., Campbell, M., & Jansen, J. O. (2020) · Journal of the Intensive Care Society · doi:10.1177/1751143720914462 · Cited by 2
A Bayesian Solution to Non-convergence of Crossed Random Effects Models
Huang, M., & Anderson, C. J. (2021) · Springer proceedings in mathematics & statistics · doi:10.1007/978-3-030-74772-5_27 · Cited by 2
A Bayesian Statistical Approach to Continuous Qualification of a Bioassay
Novick, S. J., Christian, E. A., Farmer, E., et al. (2020) · PDA Journal of Pharmaceutical Science and Technology · doi:10.5731/pdajpst.2019.011221 · Cited by 2
Confidence Interval for Coefficient of Variation of Inverse Gamma Distributions
Kaewprasert, T., Niwitpong, S., & Niwitpong, S. (2020) · Lecture notes in computer science · doi:10.1007/978-3-030-62509-2_34 · Cited by 2
Park, J., Choi, T., & Chung, Y. (2021) · Biometrical Journal · doi:10.1002/bimj.201900280 · Cited by 2
Sample Bias Related to Household Role
Hitczenko, M. (2020) · Journal of Survey Statistics and Methodology · doi:10.1093/jssam/smaa001 · Cited by 2
“Assumed priors are µ ∼ Normal(20, 7), a diffuse distribution over the general range of population means estimated from previous years’ data, and σ ∼ half-Cauchy(0, 2.5), a weakly-informative prior for between group variance (Gelman 2006; McElreath 2016).”
BackgroundInfluentialSimultaneous inference of periods and period-luminosity relations for Mira variable stars
He, S., Lin, Z., Yuan, W., et al. (2021) · The Annals of Applied Statistics · doi:10.1214/21-aoas1440 · Cited by 2
“Gelman (2006) observed that choice of “non-informative" prior distributions on hierarchical variance parameters may make a big impact on the inference, and improper choice may lead to inaccurate posterior distributions. In our numerical studies, by setting non-extreme values r̄ = 1, n̄= 1 for the variance parameters,…”
BackgroundMethodInfluentialSpatial modelling of Lexis mortality data
Divino, F., Belay, D. B., Keilman, N., & Frigessi, A. (2021) · Spatial Statistics · doi:10.1016/j.spasta.2021.100532 · Cited by 2
Uncertainty Quantification Enforced Flash Radiography Reconstruction by Two-Level Efficient MCMC
Li, Q., Xu, J., Wang, J., Yuefeng, J., & Wang, X. (2021) · IEEE Transactions on Image Processing · doi:10.1109/tip.2021.3101929 · Cited by 2
Automatic estimation of spatial spectra via smoothing splines
Zhang, S. (2021) · Computational Statistics · doi:10.1007/s00180-021-01141-z · Cited by 2
“As suggested by Gelman et al. (2014), we set the product L to 1.”
BackgroundOn the Use of Structured Prior Models for Bayesian Compressive Sensing of Modulated Signals
Marnissi, Y., Hawwari, Y., Assoumane, A., Abboud, D., & Badaoui, M. E. (2021) · Applied Sciences · doi:10.3390/app11062626 · Cited by 3
“Hence, we assign a non-informative Jeffrey prior for τ: p(τ) ∝ τ−1 [48].”
MethodShrinkage estimation of fixed and random effects in linear quantile mixed models
Ji, Y., & Shi, H. (2021) · Journal of Applied Statistics · doi:10.1080/02664763.2021.1962262 · Cited by 2
“[19] f (x|a) = 2 π(1 + (x/a)2) , x > 0.”
A Bayesian Copula Approach for Flood Analysis
Kamaruzaman, I. F., Zin, W. Z. W., & Ariff, N. M. (2021) · Malaysian Journal of Fundamental and Applied Sciences · doi:10.11113/mjfas.v17n4.2052 · Cited by 2
“This class of prior distributions is classified as weakly informative prior [12,13].”
Estimating population size and survival of adult northern pike (Esox lucius) in Lower Green Bay
Stuart, C. E., Doll, J. C., Forsythe, P. S., et al. (2022) · Fisheries Management and Ecology · doi:10.1111/fme.12534 · Cited by 2
Harezlak, J., Ruppert, D., & WAND, M. P. (2018) · Use R! · doi:10.1007/978-1-4939-8853-2_2 · Cited by 2
BAYESIAN INFERENCE FOR HIERARCHICAL AGE-PERIOD-COHORT MODELS OF REPEATED CROSS-SECTION SURVEY DATA
Yang, Y. (2006) · Sociological Methodology · doi:10.1111/j.1467-9531.2006.00003.x · Cited by 2
On Hierarchical Bayesian Spatial Small Area Model for Binary Data under Spatial Misalignment
Muchie, K. F., Wanjoya, A., & Mwalili, S. M. (2022) · Journal of Probability and Statistics · doi:10.1155/2022/3865626 · Cited by 2
Smoothing County-Level Sampling Variances to Improve Small Area Models’ Outputs
Chen, L., Sartore, L., Benecha, H. K., et al. (2022) · Stats · doi:10.3390/stats5030052 · Cited by 2
“Gelman, A.; Rubin, D.B. Inference from Iterative Simulation Using Multiple Sequences.”
MethodSoave, G. P. (2022) · Applied Economics · doi:10.1080/00036846.2022.2089347 · Cited by 2
Bayesian A-optimal two-phase designs with a single blocking factor in each phase
Vo‐Thanh, N., & Piepho, H. (2022) · Statistics and Computing · doi:10.1007/s11222-022-10126-x · Cited by 1
Sato, Y., Inomata, E., Nagoe, H., Ito, M., Konishi, T., Fujimura, H., Tanaka, A., & Nishihara, G. N. (2022) · Journal of Applied Phycology · doi:10.1007/s10811-022-02881-1 · Cited by 2
Nobre, W. S., Schmidt, A. M., Moodie, E. E. M., & Stephens, D. A. (2023) · Journal of the Royal Statistical Society Series C (Applied Statistics) · doi:10.1093/jrsssc/qlad034 · Cited by 4
“The Bayesian specification of the propensity score model is completed after assuming a diffuse normal prior for γ, a Half-Cauchy(0,1) prior for φ (Gelman, 2006); for λ, we assign a folded-normal prior whose mode provides a practical range (distance at which the correlation function drops to 5%) at half of the maximum…”
BackgroundMethodInfluentialPretorius, M., Wimberger, K., & O'Riain, M. (2023) · Forest Ecology and Management · doi:10.1016/j.foreco.2023.121078 · Cited by 1
Carter, G. M., Stolen, E. D., Breininger, D. R., Legare, S. A., Hunt, D. K., Schumann, C., & Payne, W. V. (2023) · Ornithological Applications · doi:10.1093/ornithapp/duad033 · Cited by 2
“This produced a distribution similar in shape to the Half–Cauchy distribution recommended by (Gelman 2006), but with a much shorter tail.”
MethodUnder-reported time-varying MINAR(1) process for modeling multivariate count series
Aghabazaz, Z., & Kazemi, I. (2023) · Computational Statistics & Data Analysis · doi:10.1016/j.csda.2023.107825 · Cited by 2
A Bayesian Regression Analysis of the Effects of Narcissism on Group Performance
Gunnesch‐Luca, G., Paul, K. I., & Moser, K. (2023) · Journal of Personnel Psychology · doi:10.1027/1866-5888/a000324 · Cited by 3
Naveiras, M., & Cho, S. (2023) · Applied Psychological Measurement · doi:10.1177/01466216231209758 · Cited by 4
“However, Gelman (2006) does not recommend using inverse-gamma ( ϵ , ϵ ) distributions as noninformative priors, because the resulting inferences when estimating near-zero standard deviations are highly dependent upon the choice of ϵ .”
BackgroundMethodInfluentialNunchaku: optimally partitioning data into piece-wise contiguous segments
Huo, Y., Li, H. H., Wang, X. D., Du, X., & Swain, P. S. (2023) · Bioinformatics · doi:10.1093/bioinformatics/btad688 · Cited by 1
Modeling left-censored skewed spatial processes: The case of arsenic drinking water contamination
Zhang, Q., Schmidt, A. M., & Chaubey, Y. P. (2024) · Spatial Statistics · doi:10.1016/j.spasta.2024.100816 · Cited by 2
Chen, G., & O’Malley, A. J. (2024) · Studies in computational intelligence · doi:10.1007/978-3-031-53499-7_2 · Cited by 2
Alvares, D., & Mercier, F. (2024) · Statistics in Medicine · doi:10.1002/sim.10128 · Cited by 2
Chattopadhyay, A., & Deb, S. (2024) · AStA Advances in Statistical Analysis · doi:10.1007/s10182-024-00507-0 · Cited by 2
“The usage of inverse gamma was analyzed by Gelman (2006), who commented on utilities of non-informative prior distributions for variance parameters in hierarchical models.”
Scalar Inferencing, Polarity and Cognitive Load
Marty, P., Romoli, J., Sudo, Y., van Tiel, B., & Breheny, R. (2024) · Glossa Psycholinguistics · doi:10.5070/g60112566 · Cited by 2
“3 We follow here the recommendations of Gelman (2006), Gelman, Carlin, Stern, and Rubin (2013) and McElreath (2015).”
BackgroundGridding and Parameter Expansion for Scalable Latent Gaussian Models of Spatial Multivariate Data
Peruzzi, M., Banerjee, S., Dunson, D., & Finley, A. O. (2025) · Bayesian Analysis · doi:10.1214/25-ba1515 · Cited by 9
Quirós, A., Wilson, S., Díez, R. M., et al. (2015) · Digital Signal Processing · doi:10.1016/j.dsp.2015.06.008 · Cited by 2
Genome Scanning for Quantitative Trait Loci
Xu, S. (2012) · Springer eBooks · doi:10.1007/978-0-387-70807-2_8 · Cited by 4
Paciorek, C. J., & McLACHLAN, J. S. (2008) · Collection of Biostatistics Research Archive · Cited by 2
Bayesian estimation of genotypic and phenotypic correlations from crop variety trials
Omer, S. O., Abdalla, A. W. H., Mohammed, M., et al. (2016) · Crop Breeding and Applied Biotechnology · doi:10.1590/1984-70332016v16n1a3 · Cited by 2
“The most commonly used priors for variance components in terms of the standard deviation components have been used (Gelman 2006), but classes of other relevant priors (Crossa et al. 2010) may also be included to examine support from data using the deviance information criterion.”
MethodBellera, C. A., Proust‐Lima, C., Joseph, L., et al. (2016) · Statistical Methods in Medical Research · doi:10.1177/0962280216660127 · Cited by 2
Bodnar, O., & Elster, C. (2016) · AStA Advances in Statistical Analysis · doi:10.1007/s10182-016-0279-7 · Cited by 2
Bayesian modelling of lung function data from multiple‐breath washout tests
Mahar, R. K., Carlin, J. B., Ranganathan, S., Ponsonby, A., Vuillermin, P., & Vukcevic, D. (2018) · Statistics in Medicine · doi:10.1002/sim.7650 · Cited by 2
Rojas‐Gualdrón, D. F. (2017) · Revista Brasileira de Epidemiologia · doi:10.1590/1980-5497201700030011 · Cited by 2
Martina, R., Houbiers, J. G. A., Melis, J., et al. (2018) · Biometrical Journal · doi:10.1002/bimj.201700210 · Cited by 2
Spatio-temporal quantification of climate model errors in a Bayesian framework
Arisido, M. W., Gaetan, C., Zanchettin, D., López‐Parages, J., & Rubino, A. (2018) · Stochastic Environmental Research and Risk Assessment · doi:10.1007/s00477-018-1635-9 · Cited by 2
“Here flb0 ;Rb0 ; a1; b1; a2; b2g are called hyper-parameters in the Bayesian context, and their values could either be chosen or could be assigned another priors (see, e.g., Gelman 2006).”
BackgroundHealth Effects of Sanitation Facilities: A Bayesian Semiparametric Analysis of Compositional Data
Watson, S., Sartori, J., Uthman, O. A., & Lilford, R. J. (2019) · Journal of the Royal Statistical Society Series C (Applied Statistics) · doi:10.1111/rssc.12339 · Cited by 2
“The priors for the standard deviation of these coefficients and the other hierarchical parameters are assigned half-t4-priors (Gelman, 2006), which imply a prior 95% credible interval of [0,3.5].”
BackgroundInfluentialAnders, R., van Maanen, L., & Alario, F. ‐. (2019) · Cognitive Neuropsychology · doi:10.1080/02643294.2019.1610371 · Cited by 2
“Such developments have consisted of constraining error through prior settings (Bernardo & Smith, 2000; Gelman, 2006; Gelman, Carlin, Stern, & Rubin, 2004); specifying population distributions over observed participants or items (Rouder et al.”
BackgroundBayesian modelling of lung function data from multiple-breath washout tests
Mahar, R. K., Carlin, J. B., Ranganathan, S., et al. (2018) · Figshare · Cited by 2
Lord, D., Qin, X. D., & Geedipally, S. R. (2021) · Elsevier eBooks · doi:10.1016/b978-0-12-816818-9.00009-3 · Cited by 2
Integration of statistical and administrative agricultural data from Namibia
Berg, E. J., Im, J., Zhu, Z., et al. (2021) · Statistical Journal of the IAOS · doi:10.3233/sji-200634 · Cited by 2
Ordering of Hierarchies in Hierarchical Models: Bone Mineral Density Estimation
Walsh, C., & Mengersen, K. L. (2012) · Wiley series in probability and statistics · doi:10.1002/9781118394472.ch9 · Cited by 2
Bayesian Population Modeling of Effective Connectivity.
Cosman, E. R., & Wells, W. M. (2005) · Lecture notes in computer science · doi:10.1007/11505730_4 · Cited by 2
“We subscribe to the view in [6] that “any noninformative prior distribution [is] inherently provisional— after the model has been fit, one should look at the posterior distribution to see if it makes sense.”
BackgroundKulathinal, S., Gasbarra, D., Kinra, S., Ebrahim, S. B. J., & Sillanpää, M. J. (2008) · Genetics Research · doi:10.1017/s0016672308009348 · Cited by 4
B ayesian Statistics in Quantitative Risk Assessment
Congdon, P. D. (2008) · Wiley eBooks · doi:10.1002/9780470061596.risk0518 · Cited by 2
Steps towards comprehensive Bayesian decision analysis in fisheries and environmental management
Juntunen, T. (2013) · Helda - University of Helsinki Institutional Repository · Cited by 2
Dirks, N. L. (2010) · doi:10.21007/etd.cghs.2010.0072 · Cited by 3
“Gelman in which various uninformative prior distributions for variance parameters in hierarchical models are discussed (161).”
BackgroundRisser, M. D., Calder, C. A., Berrocal, V. J., et al. (2016) · arXiv (Cornell University) · Cited by 2
SEMI-PARAMETRIC COVARIATE-MODULATED LOCAL FALSE DISCOVERY RATE FOR GENOME-WIDE ASSOCIATION STUDIESPreprint
Zablocki, R., Levine, R. A., Schork, A. J., et al. (2017) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/183384 · Cited by 2
“…conditions that have been noted among patients with schizophrenia in prior research (Greiner and Nicolson, 1965; Lidow, 2003; Battaglino et al., 2004; Leucht et al., 2007; Putnam, Sun and Zhao, 2011; Maiti et al., 2011; Buckley, Pillai and Howell, 2011; Gardiner et al., 2012; Liu et al., 2013).”
ResultA Decision Theoretic Approach to A/B TestingPreprint
Goldberg, D. E., & Johndrow, J. E. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1710.03410 · Cited by 2
“The prior on μ is a default, rather vague, prior on a location parameter, and the prior on τ is recommended as a prior on variance components in hierarchical models by Gelman [4] and Polson and Scott [12].”
BackgroundSpill, Y. G., Castillo, D., & Martı́-Renom, M. A. (2017) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/214403 · Cited by 2
“For reads aligning to the left (respectively right) of a cut site i , the number of contacts involving i are therefore made proportional to their genomic bias ι i (respectively ρ i ). ι and ρ are modelled using p-splines [30, 32, 33, 47–49].”
MethodRethinking Robust Statistics with Modern Bayesian MethodsPreprint
Williams, D. R., & Martin, S. R. (2017) · PsyArXiv (OSF Preprints) · doi:10.31234/osf.io/vaw38 · Cited by 5
A Bayesian Mark Interaction Model for Analysis of Tumor Pathology ImagesPreprint
Li, Q., Wang, X., Liang, F., et al. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1802.08308 · Cited by 2
Structure learning of Bayesian networks involving cyclic structuresPreprint
Więcek, W., Bois, F. Y., & Gayraud, G. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1906.04992 · Cited by 3
“When q > 1, the prior may be biased when the true variance is low, even with large sample sizes, as discussed by Gelman [11]; see also Alvarez et al., [3], for a simulation study.”
BackgroundAnalysis of Gaussian Spatial Models with Covariate Measurement ErrorPreprint
Tadayon, V. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1811.05648 · Cited by 2
“The use of half-normal prior for variance parameters also has been studied by Gelman (2006) and Lambert et al. (2005) Prior on ω2.”
MethodInfluentialClustering and Variable Selection in the Presence of Mixed Variable Types and Missing DataPreprint
Storlie, C. B., Myers, S. M., Katusic, S. K., et al. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1703.08741 · Cited by 2
Fuzzy Bayesian LearningPreprint
Pan, I., & Bester, D. W. (2016) · arXiv (Cornell University) · doi:10.48550/arxiv.1610.09156 · Cited by 2
Bayesian Methods for Gene Expression Analysis
Lewin, A., Bottolo, L., & Richardson, S. T. (2019) · Wiley eBooks · doi:10.1002/9781119487845.ch30 · Cited by 5
A Predictive Modeling Approach for Assessing Seismic Soil Liquefaction Potential Using CPT Data
Schmidt, J. P. (2019) · doi:10.15368/theses.2019.90 · Cited by 2
“97 the other hand, contains enough information to limit the mode to “realistic” parameter values while remaining relatively uninformative over this range (Gelman, 2006).”
BackgroundInfluentialStructured hierarchical models for probabilistic inference from perturbation screening data
Dirmeier, S., & Beerenwinkel, N. (2022) · The Annals of Applied Statistics · doi:10.1101/848234 · Cited by 4
“example, subjectively or by formal rules (Kass and Wasserman (1996), Gelman (2006), Kass and Natarajan (2006)).”
BackgroundMethodBayesian Inference for Big Spatial Data Using Non-stationary Spectral SimulationPreprint
Yang, H., & Bradley, J. R. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2001.06477 · Cited by 2
Bayesian Estimates of Transmission Line Outage Rates That Consider Line Dependencies
Zhou, K., Dobson, I., Wang, Z., et al. (2024) · IEEE Power & Energy Society General Meeting (PESGM) · doi:10.1109/pesgm51994.2024.10689155 · Cited by 2
“Gelman [14], however, does not recommend the inversegamma prior as the estimation of σ(2) would be sensitive to the parameters of inverse-gamma distribution when σ(2) is near zero.”
BackgroundBayesian Joint Models for Longitudinal and Survival Data
Armero, C. (2021) · Wiley StatsRef: Statistics Reference Online · doi:10.1002/9781118445112.stat08129 · Cited by 2
“…for θ. Priors for β, γ and α are commonly elicited as normal distributed centered at zero with wide variances, inverse Wishart for Σb and inverse gamma for σ 2 (Guo and Carlin, 2004), although options such as uniform or half-Cauchy distributions for variances seem more appropriate (Gelman, 2006).”
BackgroundA Bayesian semi-parametric model for thermal proteome profilingPreprint
Fang, S., Kirk, P., Bantscheff, M., et al. (2020) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2020.11.14.382747 · Cited by 2
Green, E. J., Finley, A. O., & Strawderman, W. E. (2020) · Springer eBooks · doi:10.1007/978-3-030-60750-0_3 · Cited by 2
Paetz, F., & Guhl, D. (2018) · SSRN Electronic Journal · doi:10.2139/ssrn.3193808 · Cited by 2
X vs. Y: An Analysis of Intergenerational Differences in Transport Mode Use Among Young AdultsPreprint
Krueger, R., Rashidi, T. H., & Vij, A. (2018) · doi:10.31235/osf.io/unezy · Cited by 2
Describing Rosé: An Embedding-Based Method for Measuring PreferencesPreprint
Mukherjee, A., & Chang, H. H. (2021) · SSRN Electronic Journal · doi:10.2139/ssrn.3859740 · Cited by 2
Swinton, P. A., Burgess, K., Hall, A., Greig, L., Psyllas, J., Aspe, R., Maughan, P., & Murphy, A. (2021) · doi:10.51224/srxiv.11 · Cited by 4
Loyalty Currency and Mental Accounting: Do Consumers Treat Points Like Money?Preprint
Lim, F., Chun, S. Y., & Satopää, V. (2021) · SSRN Electronic Journal · doi:10.2139/ssrn.3974642 · Cited by 2
Gelman, A. (2008) · Journal of the American Statistical Association · doi:10.1198/016214507000000653 · Cited by 2
Self-citationBayesian Spatial Regression for Multi-source Predictive Mapping
Finley, A. O., & Banerjee, S. (2008) · Encyclopedia of GIS · doi:10.1007/978-0-387-35973-1_97 · Cited by 2
Meta-analysis of ERP correlates of recollection and familiarityPreprint
Morcom, A. M., Kwon, S., Rugg, M. D., Curran, T., & Wiegand, R. (2022) · doi:10.31219/osf.io/f52mb · Cited by 2
Multilevel structural equation modeling for 2-1-1 mediation using Bayesian method
Kim, M. (2022) · MOspace Institutional Repository (University of Missouri) · doi:10.32469/10355/91674 · Cited by 2
Acute isolation is associated with increased reward responsiveness in human adolescentsPreprint
Tomova, L., Towner, E., Thomas, K., & Blakemore, S. (2023) · Research Square · doi:10.21203/rs.3.rs-2718114/v1 · Cited by 2
Bayesian SEM with Small Samples: Precautions and Guidelines
Smid, S. C. (2023) · Utrecht University Repository (Utrecht University) · doi:10.33540/1755 · Cited by 4
Back to Bayes-ics: Improving Universal Screening Decisions by Quantifying UncertaintyPreprint
Hall, G. J., & Doyle, E. (2023) · doi:10.35542/osf.io/rkzna · Cited by 2
The Role of Art Knowledge Training on Aesthetic Judgements and Executive FunctionsPreprint
Bara, I., Cross, E. S., & Ramsey, R. (2023) · SSRN Electronic Journal · doi:10.2139/ssrn.4514927 · Cited by 2
Changing Patterns of Risk-Sharing Channels in the United States and the Euro AreaPreprint
Cimadomo, J., Giuliodori, M., Lengyel, A., et al. (2023) · SSRN Electronic Journal · doi:10.2139/ssrn.4592980 · Cited by 2
“As discussed in Gelman (2006), this prior corresponds to a uniform prior on the standard deviation.”
BackgroundAwareness of embodiment enhances enjoyment and engages sensorimotor corticesPreprint
Moffat, R., & Cross, E. S. (2023) · doi:10.31234/osf.io/y5s89 · Cited by 2
Data analysis of cascading outages using historical data to mitigate blackouts
Zhou, K. (2022) · doi:10.31274/td-20240329-193 · Cited by 2
Aspheim, J. C. H., Aase, K., Bolstad, G., Jensen, H., & Muff, S. (2024) · bioRxiv · doi:10.1101/2024.06.01.596874 · Cited by 2
“Even though the respective Gamma prior has been criticized in other contexts (Gelman, 2006; Hodges, 2013), its use led to quite accurate results in a related study (Selle et al., 2019).”
ResultLin, C.-W. (2011) · doi:10.17077/etd.9huqp2h8 · Cited by 3
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.