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
brms : An R Package for Bayesian Multilevel Models Using Stan
Bürkner, P.-C. (2017) · Journal of Statistical Software · doi:10.18637/jss.v080.i01 · Cited by 9,476
“For other reasonable priors on standard deviations see Gelman (2006).”
BackgroundA Re-Evaluation of Random-Effects Meta-Analysis
Higgins, J. P. T., Thompson, S. G., & Spiegelhalter, D. J. (2009) · Journal of the Royal Statistical Society Series A: Statistics in Society · doi:10.1111/j.1467-985x.2008.00552.x · Cited by 2,569
“It has been suggested that a prior distribution on τ might comprise the positive part of a symmetric distribution around 0, say Cauchy or normal (Gelman, 2006).”
BackgroundPolygenic prediction via Bayesian regression and continuous shrinkage priors
Ge, T., Chen, C.-Y., Ni, Y., Feng, Y.-C. A., & Smoller, J. W. (2019) · Nature Communications · doi:10.1038/s41467-019-09718-5 · Cited by 1,922
“…on prior beliefs about the sparsity of the genetic architecture, or searching a small number of grid values, the global parameter could be learnt from data using empirical Bayes or a full Bayesian approach by placing, for example, a half-Cauchy prior on it [Gelman, 2006; Polson and Scott, 2010].”
The BUGS project: evolution, critique and future directions
Lunn, D., Spiegelhalter, D., Thomas, A., & Best, N. (2009) · Statistics in Medicine · doi:10.1002/sim.3680 · Cited by 1,876
Methods to estimate the between‐study variance and its uncertainty in meta‐analysis
Veroniki, A. A., Jackson, D., Viechtbauer, W., Bender, R., Bowden, J., Knapp, G., Kuss, O., Higgins, J. P., Langan, D., & Salanti, G. (2016) · Research Synthesis Methods · doi:10.1002/jrsm.1164 · Cited by 1,685
“An inversegamma prior with small hyper-parameters is often considered to be an approximately non-informative prior, but it was shown that inferences can be sensitive to the choice of hyper-parameters (Chung et al., 2013; Gelman, 2006).”
BackgroundWhy We (Usually) Don't Have to Worry About Multiple Comparisons
Gelman, A., Hill, J., & Yajima, M. (2012) · Journal of Research on Educational Effectiveness · doi:10.1080/19345747.2011.618213 · Cited by 1,274
“However (particularly for this kind of simple model) it is not difficult to choose weakly informative priors (Gelman, 2006).”
BackgroundSelf-citationBayesian statistics and modelling
van de Schoot, R., Depaoli, S., King, R., Kramer, B., Märtens, K., Tadesse, M. G., Vannucci, M., Gelman, A., Veen, D., Willemsen, J., et al. (2021) · Nature Reviews Methods Primers · doi:10.1038/s43586-020-00001-2 · Cited by 1,176
Self-citationPenalising Model Component Complexity: A Principled, Practical Approach to Constructing Priors
Simpson, D., Rue, H., Riebler, A., Martins, T. G., & Sørbye, S. H. (2017) · Statistical Science · doi:10.1214/16-sts576 · Cited by 1,085
“Contrary to this, Gelman (2006), Gelman et al. (2008) argue that, when setting priors on the standard deviation of a normal distribution, heavy tails are necessary for robustness (O’Hagan and Pericchi, 2012). Unfortunately, this theory is only developed for inferring location-scale families or natural parameters in ex…”
BackgroundInfluentialSavović, J., Jones, H. E., Altman, D. G., Harris, R. J., Jüni, P., Pildal, J., Als-Nielsen, B., Balk, E. M., Gluud, C., Gluud, L. L., et al. (2012) · Annals of Internal Medicine · doi:10.7326/0003-4819-157-6-201209180-00537 · Cited by 1,049
Wang, Y., Zhou, S., Yang, F., Qi, X., Wang, X., Guan, X., Shen, C., Duma, N., Vera Aguilera, J., Chintakuntlawar, A., et al. (2019) · JAMA Oncology · doi:10.1001/jamaoncol.2019.0393 · Cited by 862
Inferring change points in the spread of COVID-19 reveals the effectiveness of interventions
Dehning, J., Zierenberg, J., Spitzner, F. P., Wibral, M., Neto, J. P., Wilczek, M., & Priesemann, V. (2020) · Science · doi:10.1126/science.abb9789 · Cited by 819
How Many Countries for Multilevel Modeling? A Comparison of Frequentist and Bayesian Approaches
Stegmueller, D. (2013) · American Journal of Political Science · doi:10.1111/ajps.12001 · Cited by 818
“More sophisticated priors have been proposed (Natarajan and Kass 2000; Gelman 2006); however, I employ specifications which are commonly used by researchers (e.g. Spiegelhalter et al. 1997; Gelman and Hill 2007).”
BackgroundMethodInfluentialA review of Bayesian variable selection methods: what, how and which
O'Hara, R. B., & Sillanpää, M. J. (2009) · Bayesian Analysis · doi:10.1214/09-ba403 · Cited by 771
“σβ ∼ U(0, 20), (5) where U(a, b) denotes a uniform distribution between a and b (for a justification of this prior, see Gelman (2006)).”
BackgroundMethodInfluentialDenwood, M. J. (2016) · Journal of Statistical Software · doi:10.18637/jss.v071.i09 · Cited by 731
“The info variable in this code contains the name and directory of the simulation, which is given to the user if the object is printed.”
The Effect of Small Sample Size on Two-Level Model Estimates: A Review and Illustration
McNeish, D. M., & Stapleton, L. M. (2016) · Educational Psychology Review · doi:10.1007/s10648-014-9287-x · Cited by 715
“Although not in the context of a simulation, Gelman (2006) used Bayesian MCMC to reasonably estimate a model for data with less than five clusters.”
BackgroundMethodInfluentialRiebler, A., & Held, L. (2017) · Biometrical Journal · doi:10.1002/bimj.201500263 · Cited by 624
“Alternatives include half-normal distributions for the standard deviations (Gelman, 2006) and the recently proposed penalized complexity (PC) priors (Simpson et al., 2015).”
BackgroundLemoine, N. P. (2019) · Oikos · doi:10.1111/oik.05985 · Cited by 610
“An noninformative prior on σB with large ν yields a standard mixed-effects model with random parameters (Gelman 2006).”
BackgroundMethodInfluentialA comparison of Bayesian and likelihood-based methods for fitting multilevel models
Browne, W. J., & Draper, D. (2006) · Bayesian Analysis · doi:10.1214/06-ba117 · Cited by 603
Multiple sparse priors for the M/EEG inverse problem.
Friston, K., Harrison, L., Daunizeau, J., Kiebel, S., Phillips, C., Trujillo-Barreto, N., Henson, R., Flandin, G., & Mattout, J. (2008) · NeuroImage · doi:10.1016/j.neuroimage.2007.09.048 · Cited by 589
“Strictly speaking, this is a weakly informative hyperprior, as characterized by Gelman (2006); i.e., a weakly informative proper distribution “ that is set up so that the information it does provide is intentionally weaker than whatever actual prior knowledge is available ” .”
Struggles with Survey Weighting and Regression Modeling
Gelman, A. (2007) · Statistical Science · doi:10.1214/088342306000000691 · Cited by 604
Self-citationGordon, B. A., Blazey, T. M., Su, Y., Hari-Raj, A., Dincer, A., Flores, S., Christensen, J., McDade, E., Wang, G., Xiong, C., et al. (2018) · The Lancet Neurology · doi:10.1016/s1474-4422(18)30028-0 · Cited by 602
On Using Bayesian Methods to Address Small Sample Problems
McNeish, D. (2016) · Structural Equation Modeling: A Multidisciplinary Journal · doi:10.1080/10705511.2016.1186549 · Cited by 570
“As Gelman (2006) noted and as echoed in simulation work by McNeish (2016), with small samples, the idea of noninformative priors is more myth than reality: when the information contained in the likelihood is relatively small due to a limited sample size, the prior will necessarily play a key role in the posterior dist…”
BackgroundMethodInfluentialMultilevel (Hierarchical) Modeling: What It Can and Cannot Do
Gelman, A. (2006) · Technometrics · doi:10.1198/004017005000000661 · Cited by 557
“assuming a uniform prior distribution on γ, σy, σα, which is reasonable given that the number of counties J is large (Gelman, 2005).”
BackgroundSelf-citationA Nondegenerate Penalized Likelihood Estimator for Variance Parameters in Multilevel Models
Chung, Y., Rabe-Hesketh, S., Dorie, V., Gelman, A., & Liu, J. (2013) · Psychometrika · doi:10.1007/s11336-013-9328-2 · Cited by 554
Self-citationComparative studies of quantitative trait and neutral marker divergence: a meta‐analysis
LEINONEN, T., O’HARA, R. B., CANO, J. M., & MERILÄ, J. (2008) · Journal of Evolutionary Biology · doi:10.1111/j.1420-9101.2007.01445.x · Cited by 526
“…; b Q 0 ; b Q F ; b Q P ; l F ; lQ Nð0; 102Þ fF ; fQ expð20Þ r2d ; r 2 F ; r 2 m; r 2 Q; r 2 Tr; r 2 Tx; s 2 d ; s 2 m WCð1Þ where exp(k) denotes an exponential distribution with mean k, and WC(s) denotes a wrapped Cauchy distribution with scale s [see Gelman (2006) for a motivation of this prior].”
BackgroundA Bayesian approach for estimating length-weight relationships in fishes
Froese, R., Thorson, J. T., & Reyes, R. B. (2014) · Journal of Applied Ichthyology · doi:10.1111/jai.12299 · Cited by 523
“For the measurement error in length and weight an uninformative prior was assumed (Gelman, 2006).”
MethodUse of historical control data for assessing treatment effects in clinical trials
Viele, K., Berry, S., Neuenschwander, B., Amzal, B., Chen, F., Enas, N., Hobbs, B., Ibrahim, J. G., Kinnersley, N., Lindborg, S., et al. (2014) · Pharmaceutical Statistics · doi:10.1002/pst.1589 · Cited by 512
Population Pharmacokinetics of Therapeutic Monoclonal Antibodies
Dirks, N. L., & Meibohm, B. (2010) · Clinical Pharmacokinetics · doi:10.2165/11535960-000000000-00000 · Cited by 500
Robust meta‐analytic‐predictive priors in clinical trials with historical control information
Schmidli, H., Gsteiger, S., Roychoudhury, S., O'Hagan, A., Spiegelhalter, D., & Neuenschwander, B. (2014) · Biometrics · doi:10.1111/biom.12242 · Cited by 482
“For the case where only one historical study is available, the challenging problem of a Bayesian meta-analysis with only two studies (one historical, and one new) has to be addressed, in particular the choice of an appropriate informative prior distribution for τ (Gelman, 2006).”
BackgroundSparsity information and regularization in the horseshoe and other shrinkage priors
Piironen, J., & Vehtari, A. (2017) · Electronic Journal of Statistics · doi:10.1214/17-ejs1337si · Cited by 457
On the Half-Cauchy Prior for a Global Scale Parameter
Polson, N. G., & Scott, J. G. (2012) · Bayesian Analysis · doi:10.1214/12-ba730 · Cited by 454
The Prior Can Often Only Be Understood in the Context of the Likelihood
Gelman, A., Simpson, D., & Betancourt, M. (2017) · Entropy · doi:10.3390/e19100555 · Cited by 454
“The arguments of Simpson et al. (2017), which are inspired greatly by Gelman (2006) and Gelman et al. (2008), argue that p(σ) should have finite, non-zero density at σ = 0.”
BackgroundSelf-citationImproving transparency and replication in Bayesian statistics: The WAMBS-Checklist.
Depaoli, S., & van de Schoot, R. (2017) · Psychological Methods · doi:10.1037/met0000065 · Cited by 448
The ANTsX ecosystem for quantitative biological and medical imaging
Tustison, N. J., Cook, P. A., Holbrook, A. J., Johnson, H. J., Muschelli, J., Devenyi, G. A., Duda, J. T., Das, S. R., Cullen, N. C., Gillen, D. L., et al. (2021) · Scientific Reports · doi:10.1038/s41598-021-87564-6 · Cited by 440
The Prior Can Often Only Be Understood in the Context of the Likelihood
Gelman, A., Simpson, D., & Betancourt, M. (2017) · Entropy · doi:10.7916/d8v99mjx · Cited by 430
Self-citationModeling Clustered Data with Very Few Clusters
McNeish, D., & Stapleton, L. M. (2016) · Multivariate Behavioral Research · doi:10.1080/00273171.2016.1167008 · Cited by 430
“However, Gelman (2006) showed that these choices can actually be more informative than intended when the data have few clusters.”
BackgroundMethodResultInfluentialSparse Bayesian infinite factor models
Bhattacharya, A., & Dunson, D. B. (2011) · Biometrika · doi:10.1093/biomet/asr013 · Cited by 413
“To address these issues, Ghosh and Dunson [2009] use parameter expansion [Liu and Wu, 1999, Gelman, 2006] to induce a heavytailed default prior distribution on the loadings elements and propose an efficient Gibbs sampler.”
MethodDimensionality reduction and polynomial chaos acceleration of Bayesian inference in inverse problems
Marzouk, Y. M., & Najm, H. N. (2009) · Journal of Computational Physics · doi:10.1016/j.jcp.2008.11.024 · Cited by 418
“Adopting a fully Bayesian approach, we let θ be a hyperparameter endowed with a conjugate inverse gamma hyperprior, θ ∼ IG(α, β) [53, 64, 65]:”
MethodBrain size predicts problem-solving ability in mammalian carnivores
Benson-Amram, S., Dantzer, B., Stricker, G., Swanson, E. M., & Holekamp, K. E. (2016) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.1505913113 · Cited by 393
Summarizing historical information on controls in clinical trials
Neuenschwander, B., Capkun-Niggli, G., Branson, M., & Spiegelhalter, D. J. (2010) · Clinical Trials · doi:10.1177/1740774509356002 · Cited by 396
Ahn, W.-Y., Haines, N., & Zhang, L. (2017) · Computational Psychiatry · doi:10.1162/cpsy_a_00002 · Cited by 362
“For the group-level standard deviations, we used half-Cauchy prior distributions, which tend to give sharper and more reasonable estimates than uniform or inverse-Gaussian prior distributions (Gelman, 2006).”
MethodThe co-creation experience from the customer perspective: its measurement and determinants
Verleye, K. (2015) · Journal of Service Management · doi:10.1108/josm-09-2014-0254 · Cited by 362
“In line with the recommendations of Gelman (2006), we also checked the posterior distributions, which showed that the noninformative prior distributions made sense.”
blavaan : Bayesian Structural Equation Models via Parameter Expansion
Merkle, E. C., & Rosseel, Y. (2018) · Journal of Statistical Software · doi:10.18637/jss.v085.i04 · Cited by 360
Precipitation drives global variation in natural selection
Siepielski, A. M., Morrissey, M. B., Buoro, M., Carlson, S. M., Caruso, C. M., Clegg, S. M., Coulson, T., DiBattista, J., Gotanda, K. M., Francis, C. D., et al. (2017) · Science · doi:10.1126/science.aag2773 · Cited by 357
“, the overall intercept and average slope), parameter-expanded priors (36, 37) for the random (co)variance components (i.”
BackgroundSeroprevalence of anti–SARS-CoV-2 IgG antibodies in Kenyan blood donors
Uyoga, S., Adetifa, I. M. O., Karanja, H. K., Nyagwange, J., Tuju, J., Wanjiku, P., Aman, R., Mwangangi, M., Amoth, P., Kasera, K., et al. (2021) · Science · doi:10.1126/science.abe1916 · Cited by 328
What to believe: Bayesian methods for data analysis
Kruschke, J. K. (2010) · Trends in Cognitive Sciences · doi:10.1016/j.tics.2010.05.001 · Cited by 342
Bobak, C. A., Barr, P. J., & O’Malley, A. J. (2018) · BMC Medical Research Methodology · doi:10.1186/s12874-018-0550-6 · Cited by 344
“Although the inverse-Gamma prior distributions assumed here for the variance parameters have been shown to yield undesirable results in some applications [33], we found that they were well suited to our case study in the sense that the results were quite robust to the prior distribution parameters.”
ResultBayesian LASSO for Quantitative Trait Loci Mapping
Yi, N., & Xu, S. (2008) · Genetics · doi:10.1534/genetics.107.085589 · Cited by 344
“The former prior results in a Bayesian LASSO model (Tibshirani 1996; Yuan and Lin 2005; Park and Casella 2007), and the latter includes Jeffrey’s noninformative prior as a special case (Gelman 2006).”
MethodHow to compare cross-lagged associations in a multilevel autoregressive model.
Schuurman, N. K., Ferrer, E., de Boer-Sonnenschein, M., & Hamaker, E. L. (2016) · Psychological Methods · doi:10.1037/met0000062 · Cited by 315
Axelrad, D. A., Bellinger, D. C., Ryan, L. M., & Woodruff, T. J. (2007) · Environmental Health Perspectives · doi:10.1289/ehp.9303 · Cited by 320
“Gelman (2006) argues that more stable results can be obtained by specifying priors directly on the variance components or their square root.”
BackgroundInfluentialLeaders and International Conflict
Chiozza, G., & Goemans, H. E. (2011) · Cambridge University Press eBooks · doi:10.1017/cbo9780511996429 · Cited by 308
Bayesian Random-Effects Meta-Analysis Using the bayesmeta R Package
Röver, C. (2020) · Journal of Statistical Software · doi:10.18637/jss.v093.i06 · Cited by 310
Hahn, P. R., Murray, J. S., & Carvalho, C. M. (2020) · Bayesian Analysis · doi:10.1214/19-ba1195 · Cited by 304
Local conditions magnify coral loss after marine heatwaves
Donovan, M. K., Burkepile, D. E., Kratochwill, C., Shlesinger, T., Sully, S., Oliver, T. A., Hodgson, G., Freiwald, J., & van Woesik, R. (2021) · Science · doi:10.1126/science.abd9464 · Cited by 289
Bayesian inference for generalized linear mixed models
Fong, Y., Rue, H., & Wakefield, J. (2010) · Biostatistics · doi:10.1093/biostatistics/kxp053 · Cited by 284
“More general discussion can be found in Natarajan and Kass (2000) and Gelman (2006).”
BackgroundHierarchical Bayesian model updating for structural identification
Behmanesh, I., Moaveni, B., Lombaert, G., & Papadimitriou, C. (2015) · Mechanical Systems and Signal Processing · doi:10.1016/j.ymssp.2015.03.026 · Cited by 286
“In this case, the objective function of Equation (19) will be simplified to Equation (35) and the conditional probability distribution of 2 e can be written as Equation (36) if its prior is assumed to be uniform [70]:”
BackgroundChoy, S. L., O'Leary, R., & Mengersen, K. (2009) · Ecology · doi:10.1890/07-1886.1 · Cited by 283
Europe-Wide Dampening of Population Cycles in Keystone Herbivores
Cornulier, T., Yoccoz, N. G., Bretagnolle, V., Brommer, J. E., Butet, A., Ecke, F., Elston, D. A., Framstad, E., Henttonen, H., Hörnfeldt, B., et al. (2013) · Science · doi:10.1126/science.1228992 · Cited by 286
A comparison of conditional autoregressive models used in Bayesian disease mapping
Lee, D. (2011) · Spatial and Spatio-temporal Epidemiology · doi:10.1016/j.sste.2011.03.001 · Cited by 281
Smid, S. C., McNeish, D., Miočević, M., & van de Schoot, R. (2020) · Structural Equation Modeling: A Multidisciplinary Journal · doi:10.1080/10705511.2019.1577140 · Cited by 280
“For a more in-depth discussion of the half-Cauchy prior distribution, we refer to Gelman (2006) and Polson and Scott (2012).”
InfluentialMeager, R. (2019) · American Economic Journal: Applied Economics · doi:10.1257/app.20170299 · Cited by 278
A framework for meta-analysis of prediction model studies with binary and time-to-event outcomes
Debray, T. P., Damen, J. A., Riley, R. D., Snell, K., Reitsma, J. B., Hooft, L., Collins, G. S., & Moons, K. G. (2019) · Statistical Methods in Medical Research · doi:10.1177/0962280218785504 · Cited by 278
The Dynamics of State Policy Liberalism, 1936–2014
Caughey, D., & Warshaw, C. (2016) · American Journal of Political Science · doi:10.1111/ajps.12219 · Cited by 271
“For the prior on the innovation parameter ↵, we use a half-Cauchy distribution with a mean of 0 and a scale of 2.5 (Gelman 2006).”
MethodInfluentialStatistical methods for linguistic research: Foundational Ideas—Part II
Nicenboim, B., & Vasishth, S. (2016) · Language and Linguistics Compass · doi:10.1111/lnc3.12207 · Cited by 263
Bayesian Fusion of Multi-Band Images
Wei, Q., Dobigeon, N., & Tourneret, J.-Y. (2015) · IEEE Journal of Selected Topics in Signal Processing · doi:10.1109/jstsp.2015.2407855 · Cited by 267
“Other motivations for using this kin d of prior distribution can be found in [42].”
BackgroundInhaled Cannabis for Chronic Neuropathic Pain: A Meta-analysis of Individual Patient Data
Andreae, M. H., Carter, G. M., Shaparin, N., Suslov, K., Ellis, R. J., Ware, M. A., Abrams, D. I., Prasad, H., Wilsey, B., Indyk, D., et al. (2015) · The Journal of Pain · doi:10.1016/j.jpain.2015.07.009 · Cited by 266
Using Bayesian Deep Learning to Capture Uncertainty for Residential Net Load Forecasting
Sun, M., Zhang, T., Wang, Y., Strbac, G., & Kang, C. (2020) · IEEE Transactions on Power Systems · doi:10.1109/tpwrs.2019.2924294 · Cited by 262
“In the literature, a series of studies have been carried out on prior selection (e.g., [38], [39]).”
BackgroundParameter and State Model Reduction for Large-Scale Statistical Inverse Problems
Lieberman, C., Willcox, K., & Ghattas, O. (2010) · SIAM Journal on Scientific Computing · doi:10.1137/090775622 · Cited by 254
Darwin’s naturalization conundrum: dissecting taxonomic patterns of species invasions
Diez, J. M., Sullivan, J. J., Hulme, P. E., Edwards, G., & Duncan, R. P. (2008) · Ecology Letters · doi:10.1111/j.1461-0248.2008.01178.x · Cited by 262
“The overall intercept a0 and regression coefficients are assigned Normal prior distributions with mean 0 and variance 1000, and variance terms at each level are given broad uniform priors on the standard deviation following Gelman (2006) .”
BackgroundClassification of common human diseases derived from shared genetic and environmental determinants
Wang, K., Gaitsch, H., Poon, H., Cox, N. J., & Rzhetsky, A. (2017) · Nature Genetics · doi:10.1038/ng.3931 · Cited by 242
Simple Marginally Noninformative Prior Distributions for Covariance Matrices
Huang, A., & Wand, M. P. (2013) · Bayesian Analysis · doi:10.1214/13-ba815 · Cited by 251
A practical introduction to multivariate meta-analysis
Mavridis, D., & Salanti, G. (2013) · Statistical Methods in Medical Research · doi:10.1177/0962280211432219 · Cited by 255
Time lags in provision of habitat resources through revegetation
Vesk, P. A., Nolan, R., Thomson, J. R., Dorrough, J. W., & Nally, R. M. (2008) · Biological Conservation · doi:10.1016/j.biocon.2007.09.010 · Cited by 255
Semiparametric regression during 2003–2007
Ruppert, D., Wand, M. P., & Carroll, R. J. (2009) · Electronic Journal of Statistics · doi:10.1214/09-ejs525 · Cited by 250
Rodrigues, T. K., Suto, K., Nishiyama, H., Liu, J., & Kato, N. (2020) · IEEE Communications Surveys & Tutorials · doi:10.1109/comst.2019.2943405 · Cited by 245
“Some notable points: even if we have no guess/experience, we can use an uninformative prior, which is basically a generic distribution [190]; and, if a prior and its corresponding posterior have the same mathematical form, we are working with a conjugate prior [191].”
Methodde Villemereuil, P., Gimenez, O., & Doligez, B. (2013) · Methods in Ecology and Evolution · doi:10.1111/2041-210x.12011 · Cited by 247
“We showed in Appendix B that the commonly used inverseGamma prior is no longer a suitable option in this case, and we introduced the use of a v2 distribution following Gelman (2006)’s advice.”
MethodHow Much Can We Generalize From Impact Evaluations?
Vivalt, E. (2020) · Journal of the European Economic Association · doi:10.1093/jeea/jvaa019 · Cited by 237
Berry, S. M., Broglio, K. R., Groshen, S., & Berry, D. A. (2013) · Clinical Trials · doi:10.1177/1740774513497539 · Cited by 229
“Other parametric forms for this prior could also be chosen, including uniform or half-cauchy [25].”
MethodStochastic model specification search for Gaussian and partial non-Gaussian state space models
Frühwirth-Schnatter, S., & Wagner, H. (2010) · Journal of Econometrics · doi:10.1016/j.jeconom.2009.07.003 · Cited by 221
“This is in line with Gelman (2006) who came to similar conclusions for the related random-effects model.”
ResultDillon, J. V., Langmore, I., Tran, D., Brevdo, E., Vasudevan, S., Moore, D., Patton, B., Alemi, A., Hoffman, M., & Saurous, R. A. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1711.10604 · Cited by 226
“)) The half-Cauchy and half-Student t distributions are often used as “weakly informative” priors, which exhibit heavy tails, for variance parameters in hierarchical models [16].”
MethodHobbs, B. P., Sargent, D. J., & Carlin, B. P. (2012) · Bayesian Analysis · doi:10.1214/12-ba722 · Cited by 230
Shrinkage priors for Bayesian penalized regression
van Erp, S., Oberski, D. L., & Mulder, J. (2019) · Journal of Mathematical Psychology · doi:10.1016/j.jmp.2018.12.004 · Cited by 223
“Throughout this paper, we will consider the half-Cauchy prior on λ , i.e., λ ∼ half-Cauchy(0, 1), which is a robust alternative and a popular prior distribution in the Bayesian literature (see e.g., Gelman, 2006; Mulder & Pericchi, 2018; Polson & Scott, 2012).”
Multivariate Meta-Analysis: The Effect of Ignoring Within-Study Correlation
Riley, R. D. (2009) · Journal of the Royal Statistical Society Series A: Statistics in Society · doi:10.1111/j.1467-985x.2008.00593.x · Cited by 225
Measurement error caused by spatial misalignment in environmental epidemiology
Gryparis, A., Paciorek, C. J., Zeka, A., Schwartz, J., & Coull, B. A. (2009) · Biostatistics · doi:10.1093/biostatistics/kxn033 · Cited by 224
“Because the vague inverse gamma prior has some undesirable characteristics (Gelman, 2006), we also ran the simulations using Unif(0,1000) priors for the variance components.”
MethodWang, Y., & Zhao, T. (2017) · Géotechnique · doi:10.1680/jgeot.16.p.143 · Cited by 223
Piecing together the past: statistical insights into paleoclimatic reconstructions
Tingley, M. P., Craigmile, P. F., Haran, M., Li, B., Mannshardt, E., & Rajaratnam, B. (2012) · Quaternary Science Reviews · doi:10.1016/j.quascirev.2012.01.012 · Cited by 215
Grandchild care and welfare state arrangements in Europe
Igel, C., & Szydlik, M. (2011) · Journal of European Social Policy · doi:10.1177/0958928711401766 · Cited by 216
“On the contrary, controlling for the hierarchical structure of the data is strongly recommended in order to avoid biased estimates and standard errors irrespective of the number of observations per group (Clarke, 2008; Gelman, 2006).”
BackgroundModeling changing dependency structure in multivariate time series
Xuan, X., & Murphy, K. (2007) · Proceedings of the 24th international conference on Machine learning · doi:10.1145/1273496.1273629 · Cited by 214
“(Setting priors for variance parameters is known to be a delicate issue (Gelman, 2006), especially when considering model uncertainty.)”
BackgroundModeling Individual Effects in the Cormack–Jolly–Seber Model: A State–Space Formulation
Royle, J. A. (2008) · Biometrics · doi:10.1111/j.1541-0420.2007.00891.x · Cited by 214
Design-Comparable Effect Sizes in Multiple Baseline Designs: A General Modeling Framework
Pustejovsky, J. E., Hedges, L. V., & Shadish, W. R. (2014) · Journal of Educational and Behavioral Statistics · doi:10.3102/1076998614547577 · Cited by 213
“…model fitting may offer some advantage and should be investigated further, including in particular fully Bayesian methods with weakly informative priors (Baldwin & Fellingham, 2013; Gelman, 2006) or related varieties of penalized likelihood methods (Chung, Rabe-Hesketh, Dorie, Gelman, & Liu, 2013).”
BackgroundOstracism online: A social media ostracism paradigm
Wolf, W., Levordashka, A., Ruff, J. R., Kraaijeveld, S., Lueckmann, J.-M., & Williams, K. D. (2015) · Behavior Research Methods · doi:10.3758/s13428-014-0475-x · Cited by 212
“Bayesian analysis of variance In order to compare the effect of the manipulations on our dependent variables without inflating type 1 errors due to the large amount of contrasts such a comparison requires, we used a Bayesian analysis of variance (BANOVA) (Gelman, 2005, 2006; Kruschke, 2010).”
spBayes: An R Package for Univariate and Multivariate Hierarchical Point-referenced Spatial Models
Finley, A. O., Banerjee, S., & Carlin, B. P. (2007) · Journal of Statistical Software · doi:10.18637/jss.v019.i04 · Cited by 207
Brown, S., Hutton, B., Clifford, T., Coyle, D., Grima, D., Wells, G., & Cameron, C. (2014) · Systematic Reviews · doi:10.1186/2046-4053-3-110 · Cited by 209
Best practices for the human evaluation of automatically generated text
van der Lee, C., Gatt, A., van Miltenburg, E., Wubben, S., & Krahmer, E. (2019) · Proceedings of the 12th International Conference on Natural Language Generation · doi:10.18653/v1/w19-8643 · Cited by 205
“Mixed effects models capture such potential item aspects very well (e.g., Barr et al., 2013), and deserve more attention in NLG. Finally, Bayesian models are worth exploring, because they are less sensitive to the aforementioned problems with NHST (e.g., Gelman et al., 2006; Wagenmakers, 2007).”
BackgroundZhou, Y., & Dendukuri, N. (2014) · Statistics in Medicine · doi:10.1002/sim.6115 · Cited by 205
Magnitude and variation of prehistoric bird extinctions in the Pacific
Duncan, R. P., Boyer, A. G., & Blackburn, T. M. (2013) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.1216511110 · Cited by 204
Improved seismic fragility modeling from empirical data
Straub, D., & Der Kiureghian, A. (2008) · Structural Safety · doi:10.1016/j.strusafe.2007.05.004 · Cited by 202
DNA repair factor BRCA1 depletion occurs in Alzheimer brains and impairs cognitive function in mice
Suberbielle, E., Djukic, B., Evans, M., Kim, D. H., Taneja, P., Wang, X., Finucane, M., Knox, J., Ho, K., Devidze, N., et al. (2015) · Nature Communications · doi:10.1038/ncomms9897 · Cited by 202
Spatiotemporal hierarchical modelling of species richness and occupancy using camera trap data
Tobler, M. W., Zúñiga Hartley, A., Carrillo‐Percastegui, S. E., & Powell, G. V. N. (2015) · Journal of Applied Ecology · doi:10.1111/1365-2664.12399 · Cited by 199
Loane, M., Dolk, H., Bradbury, I., & a EUROCAT Working Group (2007) · Paediatric and Perinatal Epidemiology · doi:10.1111/j.1365-3016.2007.00820.x · Cited by 197
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.