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 501–600 of 4,089
Bayesian Nonparametric Weighted Sampling Inference ∗
Si, Y., Pillai, N. S., & Gelman, A. (2013) · Cited by 53
Self-citationBayes, Jeffreys, Prior Distributions and the Philosophy of Statistics
Gelman, A. (2009) · Columbia Academic Commons (Columbia University) · doi:10.7916/d8j38ztd · Cited by 53
Self-citationApplying Meta-Analytic-Predictive Priors with the R Bayesian Evidence Synthesis Tools
Weber, S., Li, Y., Seaman, J. W., Kakizume, T., & Schmidli, H. (2021) · Journal of Statistical Software · doi:10.18637/jss.v100.i19 · Cited by 92
A Gentle Introduction to Bayesian Network Meta-Analysis Using an Automated R Package
Liu, Y., Béliveau, A., Wei, Y., Chen, M. Y., Record-Lemon, R., Kuo, P., Pritchard, E., Tang, X., & Chen, G. (2022) · Multivariate Behavioral Research · doi:10.1080/00273171.2022.2115965 · Cited by 63
“First, Bayesian methods have been recommended over the frequentist approach for small sample sizes and complex models (Baldwin & Fellingham, 2013; Browne & Draper, 2006; Gelman, 2006; Stegmueller, 2013), which often arise in practice.”
MethodAccounting for space–time interactions in index standardization models
Thorson, J. T., & Ward, E. J. (2013) · Fisheries Research · doi:10.1016/j.fishres.2013.03.012 · Cited by 55
Bayesian approach to the design and analysis of non‐inferiority trials for anti‐infective products
Gamalo, M. A., Tiwari, R. C., & LaVange, L. M. (2013) · Pharmaceutical Statistics · doi:10.1002/pst.1588 · Cited by 57
“Gelman [36] does not recommend using a uniform prior because it tends to unduly favor higher values, resulting in excessive heterogeneity.”
Terada, R., Vo, T. D., Nishihara, G. N., Keisaku, S., Shimada, S., & Kawaguchi, S. (2015) · Journal of Applied Phycology · doi:10.1007/s10811-015-0557-x · Cited by 57
“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).”
MethodMagnetic Resonance Imaging for Locating Nonpalpable Undescended Testicles: A Meta-analysis
Krishnaswami, S., Fonnesbeck, C., Penson, D. F., & McPheeters, M. L. (2013) · PEDIATRICS · doi:10.1542/peds.2013-0073 · Cited by 55
The Many Flavors of Model‐Based Meta‐Analysis: Part I—Introduction and Landmark Data
Boucher, M., & Bennetts, M. (2016) · CPT: Pharmacometrics & Systems Pharmacology · doi:10.1002/psp4.12041 · Cited by 54
Statistical approaches for conducting network meta‐analysis in drug development
Jones, B., Roger, J., Lane, P. W., Lawton, A., Fletcher, C., Cappelleri, J. C., Tate, H., Moneuse, P., & on behalf of PSI Health Technology Special Interest Group, E. S. S. (2011) · Pharmaceutical Statistics · doi:10.1002/pst.533 · Cited by 54
Augmented mixed beta regression models for periodontal proportion data
Soto, D. M. G., Bandyopadhyay, D., & Lachos, V. H. (2014) · Statistics in Medicine · doi:10.1002/sim.6179 · Cited by 55
Sparse Estimation and Uncertainty with Application to Subgroup Analysis
Ratkovic, M., & Tingley, D. H. (2017) · Political Analysis · doi:10.1017/pan.2016.14 · Cited by 51
A cost‐effectiveness threshold based on the marginal returns of cardiovascular hospital spending
van Baal, P. V., Perry-Duxbury, M., Bakx, P., Versteegh, M., van Doorslaer, E. V., & Brouwer, W. (2018) · Health Economics · doi:10.1002/hec.3831 · Cited by 52
“In our base case analysis, we assumed noninformative/vague priors for all parameters and followed the general advice by Gelman (2006): Prior distributions for all variance parameters (σm, σγ, and στ) are defined as uniform distributions between 0 and 1.”
MethodA Default Prior Distribution for Logistic and Other Regression ModelsPreprint
Gelman, A., Jakulin, A., Su, Y., & Pittau, M. G. (2007) · SSRN Electronic Journal · doi:10.2139/ssrn.1010421 · Cited by 55
Self-citationMunira, S., Sener, I. N., & Dai, B. (2020) · Accident Analysis & Prevention · doi:10.1016/j.aap.2020.105679 · Cited by 53
Spatial Modeling of Air Pollution in Studies of Its Short‐Term Health Effects
Lee, D., & Shaddick, G. (2010) · Biometrics · doi:10.1111/j.1541-0420.2009.01376.x · Cited by 61
“The variance parameters σ2 (measurement error), σ2θ (temporal), and σ 2 φ (spatial) are assigned flat priors on the standard deviation scale as suggested by Gelman (2006).”
MethodBayesian Estimation of Categorical Dynamic Factor Models
Zhang, Z. J., & Nesselroade, J. R. (2007) · Multivariate Behavioral Research · doi:10.1080/00273170701715998 · Cited by 55
“Prior elicitation perhaps plays a crucial role in Bayesian inference (e.g., Gelman, 2006; Gelman, Carlin, Stern, & Rubin, 2003), but it is yet under researched.”
BackgroundLocally Adaptive Smoothing with Markov Random Fields and Shrinkage Priors
Faulkner, J. R., & Minin, V. N. (2017) · Bayesian Analysis · doi:10.1214/17-ba1050 · Cited by 68
Regression‐based covariance functions for nonstationary spatial modeling
Risser, M. D., & Calder, C. A. (2015) · Environmetrics · doi:10.1002/env.2336 · Cited by 75
“…elements 11 and 22 will be assigned diffuse half-Cauchy (positive-only) prior distributions, p. 11/ D C 0; s21 and p. 22/ D C 0; s22 (following Gelman, 2006), where C.c; d/ is the Cauchy distribution centered at c with scale parameter d , and the correlation will be given a uniform prior…”
BackgroundMethodGarrard, G. E., Mccarthy, M. A., Vesk, P. A., Radford, J. Q., & Bennett, A. F. (2011) · Journal of Animal Ecology · doi:10.1111/j.1365-2656.2011.01891.x · Cited by 59
“As such, we used a weakly informative half-Cauchy prior for the standard deviation amongst feeding guilds, of which there were only four (Gelman 2006).”
MethodEnhanced spatial models for predicting the geographic distributions of tick-borne pathogens
Wimberly, M. C., Baer, A. D., & Yabsley, M. J. (2008) · International Journal of Health Geographics · doi:10.1186/1476-072x-7-15 · Cited by 52
Dehning, J., Zierenberg, J., Spitzner, F. P., Wibral, M., Neto, J. P., Wilczek, M., & Priesemann, V. (2020) · medRxiv · doi:10.1101/2020.04.02.20050922 · Cited by 93
Early returning long‐distance migrant males do pay a survival cost
Lerche‐Jørgensen, M., Korner‐Nievergelt, F., Tøttrup, A. P., Willemoes, M., & Thorup, K. (2018) · Ecology and Evolution · doi:10.1002/ece3.4569 · Cited by 53
“For the among‐spe‐ cies variances of the intercept, the linear and quadratic date effects, we used a folded‐t distribution with two degrees of freedom for the standard deviation (recommended by Gelman, 2006).”
MethodDealing with Technological Change: Social Policy Preferences and Institutional Context
Busemeyer, M. R., & Tober, T. (2022) · Comparative Political Studies · doi:10.1177/00104140221139381 · Cited by 67
“Following Gelman (2006), we use half-Cauchy priors by setting the variance at the limit to 1 ðV ¼ 1Þ, the belief parameter to 1 ðnu ¼ 1Þ, the prior mean to 0 ðalpha:mu ¼ 0Þ, and the scale to 25 ðalpha:V ¼ 252Þ. 7.”
Using cognitive models to combine probability estimates
Lee, M. D., & Danileiko, I. (2014) · Judgment and Decision Making · doi:10.1017/s1930297500005799 · Cited by 47
ESTIMATING UNCERTAINTY IN MULTIVARIATE RESPONSES TO SELECTION
Stinchcombe, J. R., Simonsen, A. K., & Blows, M. W. (2013) · Evolution · doi:10.1111/evo.12321 · Cited by 62
Compliance with vessel speed restrictions to protect North Atlantic right whales
Silber, G. K., Adams, J. D., & Fonnesbeck, C. (2014) · PeerJ · doi:10.7717/peerj.399 · Cited by 52
Bayesian Inference in Semiparametric Mixed Models for Longitudinal Data
Li, Y., Lin, X., & Müller, P. (2009) · Biometrics · doi:10.1111/j.1541-0420.2009.01227.x · Cited by 51
“Several authors have noted serious difficulties in using diffuse IG priors for random effect variances (Natarajan and Kass, 1998, 2000; Gelman, 2006), and showed that the resulting posterior is near improper and the variance component estimators could be excessively biased, especially when the…”
BackgroundAssessing North American influenza dynamics with a statistical SIRS model
Hooten, M. B., Anderson, J., & Waller, L. A. (2010) · Spatial and Spatio-temporal Epidemiology · doi:10.1016/j.sste.2010.03.003 · Cited by 55
“Specifically, we used a N(0, Σ β ) prior for β and a uniform on (0, γ ) for the standard deviation σ (as suggested by Gelman, 2006).”
do Valle, D. R., & Lima, J. M. T. (2014) · Malaria Journal · doi:10.1186/1475-2875-13-443 · Cited by 54
Selection, constraint, and the evolution of coloration in African starlings
Maia, R., Rubenstein, D. R., & Shawkey, M. D. (2016) · Evolution · doi:10.1111/evo.12912 · Cited by 54
“This specification induces a scaled F-distributed marginal prior on the variances, with one degree of freedom for the numerator and denominator and scaled to the square root of the variance (Gelman 2006).”
Statistical comparison of classifiers through Bayesian hierarchical modelling
Giorgio Corani, A., Benavoli, A., Demšar, J., Mangili, F., & Zaffalon, M. (2017) · Machine Learning · doi:10.1007/s10994-017-5641-9 · Cited by 72
“The posterior distributions of the standard deviation obtained with this prior are insensitive on the choice of s̄ as long as s̄ is large enough (Kruschke, 2013; Gelman, 2006).”
BackgroundMethodInfluentialLopez, M. J., Matthews, G. J., & Baumer, B. S. (2018) · The Annals of Applied Statistics · doi:10.1214/18-aoas1165 · Cited by 52
“To assess the underlying assumptions of Models IHA and CHA, including our use of the logit transform on our probability outcomes, we use posterior predictive distribution checks, as in Gelman et al. (2014). Comparisons of Models IHA and CHA are made using the Deviance Information Criterion (DIC, Spiegelhalter et al.”
BackgroundMethodInfluentialAshton, G., Prix, R., & Jones, D. I. (2017) · Physical review. D/Physical review. D. · doi:10.1103/physrevd.96.063004 · Cited by 58
Quick, M., Li, G., & Brunton‐Smith, I. (2018) · Journal of Criminal Justice · doi:10.1016/j.jcrimjus.2018.06.003 · Cited by 56
“A positive half Gaussian prior, Normal+∞ (0, 10), was assigned to the typespecific standard deviations σsk and σek (Gelman, 2006).”
MethodDynamically borrowing strength from another study through shrinkage estimation
Röver, C., & Friede, T. (2019) · Statistical Methods in Medical Research · doi:10.1177/0962280219833079 · Cited by 62
Monitoring the COVID-19 epidemic with nationwide telecommunication data
Persson, J., Parie, J. F., & Feuerriegel, S. (2021) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.2100664118 · Cited by 59
Correa-Agudelo, E., Baye, T. M., Hernández, A., Branscum, A. J., MacKinnon, N. J., & Cuadros, D. F. (2021) · International Journal of Environmental Research and Public Health · doi:10.3390/ijerph18084021 · Cited by 52
Rivot, É., Prévost, É., Cuzol, A., Baglinière, J.-L., & Parent, É. É. (2008) · Canadian Journal of Fisheries and Aquatic Sciences · doi:10.1139/f07-153 · Cited by 52
“In agreement with Gelman (2006) statement, “we follow the approach of Bernardo (1979) and consider socalled noninformative priors as reference models to be used as a standard of comparison or starting point in place of the proper, informative prior distributions that would be appropriate for a full…”
Morris, W. K., Vesk, P. A., McCarthy, M. A., Bunyavejchewin, S., & Baker, P. J. (2015) · Ecology and Evolution · doi:10.1002/ece3.1346 · Cited by 51
“Sometimes, weakly informative priors are used (e.g., Gelman 2006; Gelman et al. 2008) but in such cases, the motivation derives from statistical concerns rather than an attempt to incorporate existing ecological knowledge.”
BackgroundHybrid statistical modelling of the frequency response function of industrial robots
Nguyen, V., & Melkote, S. (2021) · Robotics and Computer-Integrated Manufacturing · doi:10.1016/j.rcim.2021.102134 · Cited by 51
A Bayesian dose–response meta-analysis model: A simulations study and application
Hamza, T., Cipriani, A., Furukawa, T. A., Egger, M., Orsini, N., & Salanti, G. (2021) · Statistical Methods in Medical Research · doi:10.1177/0962280220982643 · Cited by 51
“First, for some scenarios, the estimation can be sensitive to the prior choice [32].”
BackgroundCorrendo, A. A., Tremblay, N., Coulter, J. A., Diaz, D. A. R., Franzen, D. W., Nafziger, E. D., Prasad, P. V. V., Rosso, L. H. M., Steinke, K., Du, J., et al. (2021) · Agricultural and Forest Meteorology · doi:10.1016/j.agrformet.2021.108668 · Cited by 72
Wang, J., López‐Lozano, R., Weiss, M., et al. (2022) · Remote Sensing of Environment · doi:10.1016/j.rse.2022.113085 · Cited by 51
Kwon, H., Brown, C., Xu, K., et al. (2009) · Hydrological Sciences Journal · doi:10.1623/hysj.54.3.582 · Cited by 51
Pollinators and seed predators generate conflicting selection on Dalechampia blossoms
Pérez‐Barrales, R., Bolstad, G. H., Pélabon, C., Hansen, T. F., & Armbruster, W. S. (2013) · Oikos · doi:10.1111/j.1600-0706.2013.20780.x · Cited by 54
A Bayesian central equivalent dose model for optically stimulated luminescence dating
Combès, B., Philippe, A., Lanos, P., Mercier, N., Tribolo, C., Guérin, G., Guibert, P., & Lahaye, C. (2015) · Quaternary Geochronology · doi:10.1016/j.quageo.2015.04.001 · Cited by 59
Long-term forecasting of volcanic explosivity
Bebbington, M. (2014) · Geophysical Journal International · doi:10.1093/gji/ggu078 · Cited by 51
A hierarchical Bayesian approach for estimation of photosynthetic parameters of C3plants
PATRICK, L. D., Ogle, K., & Tissue, D. T. (2009) · Plant Cell & Environment · doi:10.1111/j.1365-3040.2009.02029.x · Cited by 63
“Folded-Cauchy (i.e. a Student’s t-distribution with one degree of freedom) densities were assigned as priors for all standard deviations (s = 1/ √t, where t is the precision parameter of interest) as suggested by Gelman (2006).”
MethodKrauss, M., Tappe, K., Schuppert, A., Kuepfer, L., & Goerlitz, L. (2015) · PLOS ONE · doi:10.1371/journal.pone.0139423 · Cited by 51
A Hierarchical Model for Lithium-Ion Battery Degradation Prediction
Xu, X., Li, Z., & Chen, N. (2015) · IEEE Transactions on Reliability · doi:10.1109/tr.2015.2451074 · Cited by 56
“where represents the inverse-gamma distribution [17], represents the multivariate normal distribution, and represents the inverse Wishart distribution [18].”
BackgroundA Bayesian Approach to Multistate Hidden Markov Models: Application to Dementia Progression
Williams, J., Storlie, C. B., Therneau, T. M., et al. (2019) · Journal of the American Statistical Association · doi:10.1080/01621459.2019.1594831 · Cited by 49
“Although commonly done in the literature (mostly due to conjugacy), we hesitate to use inverse-gamma priors on variances for reasons discussed in Gelman et al. (2006); namely, the inverse-gamma( ε , ε ) family of priors is very sensitive to the choice of ε , which does not naturally lend itself as…”
BackgroundMethodTowards activity-based exposure measures in spatial analysis of pedestrian–motor vehicle crashes
Dong, N., Meng, F., Zhang, J., Wong, S. C., & Xu, P. (2020) · Accident Analysis & Prevention · doi:10.1016/j.aap.2020.105777 · Cited by 54
Yin, Q., Li, L., Guo, X., Wu, R., Shi, B., Wang, Y., Liu, Y., Wu, S., Pan, Y., Wang, Q., et al. (2019) · Parasites & Vectors · doi:10.1186/s13071-019-3715-1 · Cited by 54
Growth-Mindset Intervention Delivered by Teachers Boosts Achievement in Early Adolescence
Porter, T., Molina, D. C., Cimpian, A., Roberts, S., Fredericks, A., Blackwell, L. S., & Trzesniewski, K. H. (2022) · Psychological Science · doi:10.1177/09567976211061109 · Cited by 75
“For the random intercepts, we used half-normal priors with the same standard deviation as that for the fixed effects (Gelman, 2006).”
MethodGhosh, P., Tang, X., Ghosh, M., & Chakrabarti, A. (2015) · Bayesian Analysis · doi:10.1214/15-ba973 · Cited by 70
Bayes, Jeffreys, Prior Distributions and the Philosophy of Statistics
Gelman, A. (2009) · Statistical Science · doi:10.1214/09-sts284d · Cited by 60
Self-citationDefault Bayesian analysis with global-local shrinkage priors
Bhadra, A., Datta, J., Polson, N. G., & Willard, B. T. (2016) · Biometrika · doi:10.1093/biomet/asw041 · Cited by 47
Simple, scalable and accurate posterior interval estimation
Li, C., Srivastava, S., & Dunson, D. B. (2017) · Biometrika · doi:10.1093/biomet/asx033 · Cited by 50
“The model assumes that where gdP denotes the generalized double Pareto shrinkage prior of [3] and Half-t is chosen to be weakly-informative [10].”
Conservative prior distributions for variance parameters in hierarchical models
Gustafson, P., Hossain, S., & MacNab, Y. C. (2006) · Canadian Journal of Statistics · doi:10.1002/cjs.5550340302 · Cited by 50
“Gelman (2005). Prior distributions for variance parameters in hierarchical models. Bayesian Analysis, to appear. A. Gelman, J. B. Carlin, H. S. Stern & D. B. Rubin (2004). Bayesian Data Analysis, 2nd. ed. Chapman & Hall/CRC, Boca Raton. W. R. Gilks & P. Wild (1992). Adaptive rejection sampling for Gibbs sampling. Appl…”
BackgroundInfluentialAn Assessment of Fecal Indicator Bacteria-Based Water Quality Standards
Gronewold, A. D., Borsuk, M. E., Wolpert, R. L., et al. (2008) · Environmental Science & Technology · doi:10.1021/es703144k · Cited by 50
Hierarchical multi‐population viability analysis
Leasure, D. R., Wenger, S. J., Chelgren, N. D., Neville, H. M., Dauwalter, D. C., Bjork, R. D., Fesenmyer, K. A., Dunham, J. B., Peacock, M. M., Luce, C. H., et al. (2018) · Ecology · doi:10.1002/ecy.2538 · Cited by 49
“…can be modeled hierarchically with hyper-parameters shared among all populations rRi Half Cauchyðlr;rrÞ lr;rr Uniformð0; 10Þ: (3) We used Half-Cauchy distributions to represent standard deviations in realized growth rates following recommendations of Gelman (2006) for hierarchical variance models.”
MethodDindar, S., Kaewunruen, S., & An, M. (2021) · Reliability Engineering & System Safety · doi:10.1016/j.ress.2021.108130 · Cited by 49
Conducting and Evaluating Multilevel Studies: Recommendations, Resources, and a Checklist
González‐Romá, V., & Hernández, A. (2022) · Organizational Research Methods · doi:10.1177/10944281211060712 · Cited by 60
Penalized wavelets: Embedding wavelets into semiparametric regression
WAND, M. P., & Ormerod, J. T. (2011) · Electronic Journal of Statistics · doi:10.1214/11-ejs652 · Cited by 47
Bayesian Demographic Accounts: Subnational Population Estimation Using Multiple Data Sources
Bryant, J., & Graham, P. J. (2013) · Bayesian Analysis · doi:10.1214/13-ba820 · Cited by 53
Liess, A., Rowe, O. F., Guo, J., Thomsson, G., & Lind, M. I. (2013) · Journal of Animal Ecology · doi:10.1111/1365-2656.12107 · Cited by 49
“0 (R Development Core Team 2011; iterations: 50 000; burn in: 5000; thinning interval: 100) and used parameter-expanded priors (centred at 0; variance of 1000), which are noninformative but have proper distributions (Gelman 2006).”
MethodInfluentialAllowing for uncertainty due to missing data in meta‐analysis—Part 2: Hierarchical models
White, I. R., Welton, N. J., Wood, A. M., Ades, A. E., & Higgins, J. P. T. (2008) · Statistics in Medicine · doi:10.1002/sim.3007 · Cited by 49
A multiregion community model for inference about geographic variation in species richness
Sutherland, C. S., Brambilla, M., Pedrini, P., & Tenan, S. (2016) · Methods in Ecology and Evolution · doi:10.1111/2041-210x.12536 · Cited by 49
“In addition, variance for the random sessions can be difficult to estimate, or can be overestimated, when the number of sessions is small or when there are species detected in a few sessions (Gelman 2006; Tobler et al. 2015).”
BackgroundCovariate‐adjusted borrowing of historical control data in randomized clinical trials
Han, B., Zhan, J., Zhong, Z. J., Liu, D., & Lindborg, S. R. (2017) · Pharmaceutical Statistics · doi:10.1002/pst.1815 · Cited by 59
Temporal variation in pelagic food chain length in response to environmental change
Ruiz-Cooley, R. I., Gerrodette, T., Fiedler, P. C., Chivers, S. J., Danil, K., & Ballance, L. T. (2017) · Science Advances · doi:10.1126/sciadv.1701140 · Cited by 50
Bayesian estimation of mixed logit models: Selecting an appropriate prior for the covariance matrix
Akinc, D., & Vandebroek, M. L. (2017) · Journal of Choice Modelling · doi:10.1016/j.jocm.2017.11.004 · Cited by 48
Highest plasticity of carbon‐concentrating mechanisms in earliest evolved phytoplankton
de Waal, D. B. V., Brandenburg, K. M., Keuskamp, J. A., Trimborn, S., Rokitta, S. D., Kranz, S. A., & Rost, B. (2019) · Limnology and Oceanography Letters · doi:10.1002/lol2.10102 · Cited by 49
“As random factors were close to zero, we used weakly informative Cauchy priors as recommended by Gelman (2006).”
MethodDavis, K. P., Augustine, D. J., Monroe, A. P., Derner, J., & Aldridge, C. L. (2019) · Ecological Applications · doi:10.1002/eap.2020 · Cited by 53
“Due to convergence issues in preliminary analyses, we specified a weakly informative half-Cauchy prior for the standard deviation of the random term for year (Gelman 2006).”
MethodViolent crime in San Antonio, Texas: An application of spatial epidemiological methods
Sparks, C. S. (2011) · Spatial and Spatio-temporal Epidemiology · doi:10.1016/j.sste.2011.10.001 · Cited by 49
Modeling and estimation for self-exciting spatio-temporal models of terrorist activity
Clark, N., & Dixon, P. M. (2018) · The Annals of Applied Statistics · doi:10.1214/17-aoas1112 · Cited by 53
“The choice of the Half-Cauchy is in line with the recommendations for vague priors for variance components of hierarchical models as outlined in Gelman et al. (2006) and rigorously defended in Polson et al.”
MethodInfluentialEmployment Insecurity, Incumbent Partisanship, and Voting Behavior in Comparative Perspective
Helgason, A. F., & Mérola, V. (2016) · Comparative Political Studies · doi:10.1177/0010414016679176 · Cited by 49
“First, following Gelman (2006), a uniform distribution was specified on the standard deviation of the various priors.”
MethodWiksten, A. S., Hawkins, N., Piepho, H., & Gsteiger, S. (2020) · Value in Health · doi:10.1016/j.jval.2020.03.010 · Cited by 54
Ensuring identifiability in hierarchical mixed effects Bayesian models
Ogle, K., & Barber, J. J. (2020) · Ecological Applications · doi:10.1002/eap.2159 · Cited by 59
Application of Bayesian methods to accelerate rare disease drug development: scopes and hurdles
Kidwell, K., Roychoudhury, S., Wendelberger, B. A., Scott, J., Moroz, T., Yin, S., Majumder, M., Zhong, J., Huml, R. A., & Miller, V. (2022) · Orphanet Journal of Rare Diseases · doi:10.1186/s13023-022-02342-5 · Cited by 61
Specification of Random Effects in Multilevel Models: A Review
Grilli, L., & Rampichini, C. (2014) · Quality & Quantity · doi:10.1007/s11135-014-0060-5 · Cited by 62
“Gelman (2006) pointed out some drawbacks of the Gamma prior and suggested an alternative prior distribution based on the half-t family.”
BackgroundMeasurement errors and scaling relations in astrophysics: a review
Andreon, S., & Hurn, M. (2012) · Statistical Analysis and Data Mining The ASA Data Science Journal · doi:10.1002/sam.11173 · Cited by 68
IsoWeb: A Bayesian Isotope Mixing Model for Diet Analysis of the Whole Food Web
Kadoya, T., Osada, Y., & Takimoto, G. (2012) · PLoS ONE · doi:10.1371/journal.pone.0041057 · Cited by 56
Using multilevel spatial models to understand salamander site occupancy patterns after wildfire
Chelgren, N. D., Adams, M. J., Bailey, L. L., & Bury, R. B. (2011) · Ecology · doi:10.1890/10-0322.1 · Cited by 52
Waldmann, P., Hallander, J., Hoti, F. J., & Sillanpää, M. J. (2008) · Genetics · doi:10.1534/genetics.107.084160 · Cited by 47
“Gelman (2006) provides a useful discussion of alternative priors.”
BackgroundA Space–Time Multivariate Bayesian Model to Analyse Road Traffic Accidents by Severity
Boulieri, A., Liverani, S., de Hoogh, K., & Blangiardo, M. (2016) · Journal of the Royal Statistical Society Series A (Statistics in Society) · doi:10.1111/rssa.12178 · Cited by 55
“…entries to 0.0005, following Kass and Natarajan (2006), who suggested considering the data on the prior specification in the multivariate case, and Gelman (2006), who recommended taking considerable care with the prior specification on unobserved parameters and assigning reasonable values in…”
BackgroundMethodInfluentialPlanning sample sizes when effect sizes are uncertain: The power-calibrated effect size approach.
McShane, B. B., & Böckenholt, U. (2016) · Psychological Methods · doi:10.1037/met0000036 · Cited by 48
Quinn, T. P., Cunningham, C. J., & Wirsing, A. J. (2016) · Oecologia · doi:10.1007/s00442-016-3782-3 · Cited by 56
“…hierarchical prior on αs (1)P̂s,y = αsNs,y 1+ αsβ −1 s Ns,y (2) α̂s = exp(µα + να × depths) β̂s = exp(µβ + νβ × depths) (3) αs ∼ N(α̂s, σα) βs ∼ N(β̂s, σβ) (4)− ln(Ls) = Y ln(σs)+ 1 2σ 2s Y∑ y=1 (Ps,y − P̂s,y) 2 1 3 and βs: σα ∼ U(0, 10) and σβ ∼ U(0, 104) given recommendations by Gelman (2006).”
Kotilainen, M., Vanhatalo, J. P., Suominen, M., et al. (2016) · Cold Regions Science and Technology · doi:10.1016/j.coldregions.2016.12.006 · Cited by 48
Correlation of neuter status and expression of heritable disorders
Belanger, J. M., Bellumori, T. P., Bannasch, D. L., Famula, T. R., & Oberbauer, A. M. (2017) · Canine Genetics and Epidemiology · doi:10.1186/s40575-017-0044-6 · Cited by 75
“small, the weakly informative prior of a half-Cauchy for σ(2)s and σ(2)b in our hierarchical model can bring some valuable stability to our estimates of pij and the subsequent OR values [34].”
BackgroundOn the Use of Mixed Markov Models for Intensive Longitudinal Data
de Haan-Rietdijk, S., Kuppens, P., Bergeman, C. S., Sheeber, L., Allen, N. B., & Hamaker, E. L. (2017) · Multivariate Behavioral Research · doi:10.1080/00273171.2017.1370364 · Cited by 52
“We centered the predictor depression to have a mean of 0, but did not rescale it, in line with the advice of Gelman and colleagues (2008) concerning binary predictors.”
MethodResultThe Matrix-F Prior for Estimating and Testing Covariance Matrices
Mulder, J., & Pericchi, L. R. (2018) · Bayesian Analysis · doi:10.1214/17-ba1092 · Cited by 45
“Recently the half-t prior is becoming a popular alternative for a standard deviation (Gelman, 2006; Polson and Scott, 2012).”
BackgroundMethodInfluentialBradley, J. R., Holan, S. H., & Wikle, C. K. (2019) · Journal of the American Statistical Association · doi:10.1080/01621459.2019.1677471 · Cited by 56
“Additionally, extensions of the standard inverse-gamma priors for variances of a normal random variable (Gelman, 2006) are discussed in context of the LCM.”
MethodHeritability of short‐scale natal dispersal in a large‐scale foraging bird, the wandering albatross
Charmantier, A., Buoro, M., Giménez, O., & Weimerskirch, H. (2011) · Journal of Evolutionary Biology · doi:10.1111/j.1420-9101.2011.02281.x · Cited by 56
Weitkamp, L. A., Teel, D. J., Liermann, M., Hinton, S. A., Van Doornik, D. M., & Bentley, P. J. (2015) · Marine and Coastal Fisheries · doi:10.1080/19425120.2015.1047476 · Cited by 47
“For example, X stock X year .myear; stock; hat/D 0 0 : This is equivalent to treating the categories as fixed effects or the groups as a finite population (Gelman 2006), where one is interested in making comparisons between specific levels (e.g., between the lower and upper Columbia River stocks).”
BackgroundAccounting for vessel effects when standardizing catch rates from cooperative surveys
Thorson, J. T., & Ward, E. J. (2014) · Fisheries Research · doi:10.1016/j.fishres.2014.02.036 · Cited by 47
Monitoring bluetongue disease (BTV-1) epidemic in southern Spain during 2007
Allepuz, A., García‐Bocanegra, I., Napp, S., et al. (2010) · Preventive Veterinary Medicine · doi:10.1016/j.prevetmed.2010.06.005 · Cited by 47
Eagle, A. J., Eagle, D. E., Stobbe, T. E., & van Kooten, G. C. (2014) · American Journal of Agricultural Economics · doi:10.1093/ajae/aau098 · Cited by 58
“The hierarchical standard deviation parameters are assigned uniform (0,1) priors in accordance with Gelman (2006).”
A General Bayesian Approach to Analyzing Diallel Crosses of Inbred Strains
Lenarcic, A., Svenson, K. L., Churchill, G. A., & Valdar, W. S. J. (2012) · Genetics · doi:10.1534/genetics.111.132563 · Cited by 52
“In this case, “flat” priors (i.e., } 1, as given by nt ¼ 2, mt ¼ 0, after, e.g., Gelman 2006, or our choice of a nonzero mt) will often be preferable to the traditional Jeffreys prior [} 1/t2; mt ¼ 0, nt ¼ 0 (Jeffreys 1946)], which includes strong prior weight around t2q 5 0.”
BackgroundBayesian graphical models for regression on multiple data sets with different variables
Jackson, C. H., Best, N., & Richardson, S. T. (2008) · Biostatistics · doi:10.1093/biostatistics/kxn041 · Cited by 50
“Truncated positive N (0, 1) priors were used for σ 21 , . . . , σ 2 5 (Gelman, 2006).”
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.