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 301–400 of 4,089
Bayesian Quantile Regression for Count Data with Application to Environmental Epidemiology
Lee, D., & Neocleous, T. (2010) · Journal of the Royal Statistical Society Series C (Applied Statistics) · doi:10.1111/j.1467-9876.2010.00725.x · Cited by 82
“Therefore in this paper we follow the suggestion of Gelman (2006) and specify a flat prior on the standard deviation scale, which is essentially an improper inverse gamma distribution obtained by setting a=−0:5 and b= 0.”
MethodRast, P., Hofer, S. M., & Sparks, C. (2012) · Multivariate Behavioral Research · doi:10.1080/00273171.2012.658328 · Cited by 87
“At the same time, however, the variances are quite constrained and the SD of the random effect may shrink (Gelman, 2006).”
BackgroundMethodInfluentialCheng, Q., Zhang, X., Chen, L., & Liu, J. (2022) · Nature Communications · doi:10.1038/s41467-022-34164-1 · Cited by 105
Modeling Site Effects in the Design and Analysis of Multi-site Trials
Feaster, D. J., Mikulich‐Gilbertson, S. K., & Brincks, A. M. (2011) · The American Journal of Drug and Alcohol Abuse · doi:10.3109/00952990.2011.600386 · Cited by 89
Conservation Efforts May Increase Malaria Burden in the Brazilian Amazon
Valle, D., & Clark, J. (2013) · PLoS ONE · doi:10.1371/journal.pone.0057519 · Cited by 82
Bayesian hierarchical modeling based on multisource exchangeability
Kaizer, A. M., Koopmeiners, J. S., & Hobbs, B. P. (2017) · Biostatistics · doi:10.1093/biostatistics/kxx031 · Cited by 101
Four experiments on the relational dynamics and prosocial consequences of gratitude
Tsang, J.-A. C., & Martin, S. R. (2017) · The Journal of Positive Psychology · doi:10.1080/17439760.2017.1388435 · Cited by 100
“All residuals are assigned an a priori distribution of ~ Cauchy(0,1)+, following the recommendations of Gelman (2006).”
BackgroundA Hierarchical Modelling Framework for Identifying Unusual Performance in Health Care Providers
Ohlssen, D. I., Sharples, L. D., & Spiegelhalter, D. J. (2007) · Journal of the Royal Statistical Society Series A (Statistics in Society) · doi:10.1111/j.1467-985x.2007.00487.x · Cited by 82
Bayesian Inference for the Spatial Random Effects Model
Kang, E. L., & Cressie, N. (2011) · Journal of the American Statistical Association · doi:10.1198/jasa.2011.tm09680 · Cited by 94
“Gelman (2006) noticed sensitivity of inference to the choice of a and suggested a (noninformative) uniform prior density, Unif[0, κξ ], on the standard-deviation parameter σξ .”
BackgroundMethodHHV-8 seroprevalence: a global view
Rohner, E., Wyss, N., Trelle, S., Mbulaiteye, S. M., Egger, M., Novak, U., Zwahlen, M., & Bohlius, J. (2014) · Systematic Reviews · doi:10.1186/2046-4053-3-11 · Cited by 82
Quantifying Sources of Uncertainty in Projections of Future Climate*
Northrop, P. J., & Chandler, R. E. (2014) · Journal of Climate · doi:10.1175/jcli-d-14-00265.1 · Cited by 90
“Gelman (2006) considers what kind of prior distribution should be placed on a variance component s when the available data provide only limited information about it: in the present context, this is the situation for sS because data are available for just three emissions scenarios.”
BackgroundInfluentialChang, H. H., Peng, R. D., & Dominici, F. (2011) · Biostatistics · doi:10.1093/biostatistics/kxr002 · Cited by 86
Time for a change: a tutorial for comparing multiple classifiers through Bayesian analysis
Benavoli, A., Giorgio Corani, A., Demšar, J., & Marco Zaffalon, A. (2016) · arXiv (Cornell University) · doi:10.48550/arxiv.1606.04316 · Cited by 81
“s way it realistically represents the fact the estimates referring to dierent data sets data sets have dierent uncertainty. The high-level distribution of the ˙ i’s is unif(0;˙), as recommended by Gelman (2006), as it yields inferences which are insensitive to ˙, if ˙is large enough. To this end we set ˙= 1000 s (Krus…”
BackgroundPredictively consistent prior effective sample sizes
Neuenschwander, B. E., Weber, S., Schmidli, H., & O’Hagan, A. (2020) · Biometrics · doi:10.1111/biom.13252 · Cited by 105
“In the following, we use a 𝑁(0, 102) prior for 𝜇 and a half-normal prior with scale 1 for 𝜏. Alternatively, t-priors for 𝜇 and half-Cauchy or half-t pri- ors for 𝜏 could be used (Gelman, 2006; Polson and Scott, 2012).”
MethodEarly-life environmental enrichment generates persistent individualized behavior in mice
Zocher, S., Schilling, S., Grzyb, A. N., Adusumilli, V. S., Lopes, J. B., Günther, S., Overall, R. W., Winter, Y., & Kempermann, G. (2020) · Science Advances · doi:10.1126/sciadv.abb1478 · Cited by 93
“To estimate interindividual variances and covariances, we used parameter-expanded proper priors (45).”
MegaLMM: Mega-scale linear mixed models for genomic predictions with thousands of traits
Runcie, D. E., Qu, J., Cheng, H., & Crawford, L. (2021) · Genome Biology · doi:10.1186/s13059-021-02416-w · Cited by 80
Wagner, T., Soranno, P. A., Webster, K. E., & Cheruvelil, K. S. (2011) · Freshwater Biology · doi:10.1111/j.1365-2427.2011.02621.x · Cited by 88
“We used non-informative normal priors for la and lb and non-informative uniform priors for ry, ra and rb (Gelman, 2006).”
MethodVariational Bayesian Inference for Parametric and Nonparametric Regression With Missing Data
Faes, C., Ormerod, J. T., & WAND, M. P. (2011) · Journal of the American Statistical Association · doi:10.1198/jasa.2011.tm10301 · Cited by 80
Using Redundant Parameterizations to Fit Hierarchical Models
Gelman, A., van Dyk, D. A., Huang, Z., & Boscardin, W. J. (2008) · Journal of Computational and Graphical Statistics · doi:10.1198/106186008x287337 · Cited by 112
“…BETWEEN THE PARAMETER-EXPANDED MODEL AND THE USUAL MODEL The expanded model can be viewed as a generalization of the usual hierarchical regression model on β, expanding the class of prior distributions on the variance parameter σβ (as described by Gelman (2006) for the one-variance-parameter case).”
BackgroundMethodInfluentialSelf-citationDo ship strikes threaten the recovery of endangered eastern North Pacific blue whales?
Monnahan, C. C., Branch, T. A., & Punt, A. E. (2014) · Marine Mammal Science · doi:10.1111/mms.12157 · Cited by 80
Detecting ecological responses to flow variation using Bayesian hierarchical models
Webb, J. A., Stewardson, M. J., & Koster, W. M. (2009) · Freshwater Biology · doi:10.1111/j.1365-2427.2009.02205.x · Cited by 91
“Gelman (2006) describes the use of a ‘half-Cauchy’”
BackgroundMethodInfluentialLarge-network travel time distribution estimation for ambulances
Westgate, B. S., Woodard, D. B., Matteson, D. S., & Henderson, S. G. (2016) · European Journal of Operational Research · doi:10.1016/j.ejor.2016.01.004 · Cited by 87
Climate, host phylogeny and the connectivity of host communities govern regional parasite assembly
Clark, N. J., Clegg, S. M., Sam, K., Goulding, W., Koane, B., & Wells, K. (2017) · Diversity and Distributions · doi:10.1111/ddi.12661 · Cited by 85
“We used a flat prior for residual variance and parameter expansion (redundant multiplicative reparameterization of the linear model) for grouping terms, which reduces dependence among parameters and improves mixing (Gelman, 2006).”
MethodOn the importance of avoiding shortcuts in applying cognitive models to hierarchical data
Boehm, U., Marsman, M., Matzke, D., & Wagenmakers, E. (2018) · Behavior Research Methods · doi:10.3758/s13428-018-1054-3 · Cited by 86
“…participant-specific sampling variance σ 2gi is drawn from a half-normal distribution with mean 0 and standard deviation λ (fourth unshaded, stochastic node from the left at the top; see Gelman, 2006; Chung et al., 2013, for a discussion of choices for prior distributions for variance parameters).”
BackgroundMethodBürkner, P.-C., Williams, D. R., Simmons, T. C., & Woolley, J. D. (2017) · Schizophrenia Bulletin · doi:10.1093/schbul/sbx053 · Cited by 78
Test of an environmental flow release in a British Columbia river: does more water mean more fish?
Bradford, M. J., Higgins, P. S., Korman, J., & Sneep, J. (2011) · Freshwater Biology · doi:10.1111/j.1365-2427.2011.02633.x · Cited by 100
Karr, J., Areshenkoff, C. N., Duggan, E., & García-Barrera, M. (2014) · Neuropsychology Review · doi:10.1007/s11065-014-9271-8 · Cited by 74
“…relatively non-informative prior for the mean μ e N 0; 1000 ð Þ Whereas, due to the small sample size, we place a more informative prior on the variance (Gelman 2006): σ 2 e Cauchy þ 0; 1 ð Þ Which keeps the variance at a reasonable level, given that there may be too few studies to give a…”
BackgroundA Bayesian framework for comparative quantitative genetics
Ovaskainen, O., Cano, J. M., & Merilä, J. (2008) · Proceedings of the Royal Society B Biological Sciences · doi:10.1098/rspb.2007.0949 · Cited by 83
“discussion on the choice of the prior, see Gelman (2006).”
BackgroundBayesian Prior Choice in IRT Estimation Using MCMC and Variational Bayes
Natesan, P., Nandakumar, R., Minka, T., & Rubright, J. D. (2016) · Frontiers in Psychology · doi:10.3389/fpsyg.2016.01422 · Cited by 110
“A relatively informative inverse gamma (1, 1) distribution was used for variance because Gelman (2006) cautioned against the use of very low values such as 0.01 and 0.001 for the gamma prior which lead to improper posteriors.”
MethodInfluentialMultilevel Models for the Analysis of Comparative Survey Data: Common Problems and Some Solutions
Schmidt-Catran, A. W., Fairbrother, M., & Andreß, H. (2019) · KZfSS Kölner Zeitschrift für Soziologie und Sozialpsychologie · doi:10.1007/s11577-019-00607-9 · Cited by 89
A hierarchical Bayesian SED model for Type Ia supernovae in the optical to near-infrared
Mandel, K. S., Thorp, S., Narayan, G., Friedman, A. S., & Avelino, A. (2022) · Monthly Notices of the Royal Astronomical Society · doi:10.1093/mnras/stab3496 · Cited by 81
McNeish, D. (2015) · Journal of Educational and Behavioral Statistics · doi:10.3102/1076998615621299 · Cited by 91
“…prior distributions for the variance components is unlikely to be an adequate solution with small sample sizes because these priors can actually be quite informative under small sample conditions and lead to considerable bias in latent factor variance estimates (Carlin & Louis, 2000; Gelman, 2006).”
BackgroundMethodInfluentialCalabrese, J. M., Brunner, J. L., & Ostfeld, R. S. (2011) · PLoS ONE · doi:10.1371/journal.pone.0029215 · Cited by 78
Estimation of daily stream water temperatures with a Bayesian regression approach
Sohrabi, M. M., Benjankar, R. M., Tonina, D., Wenger, S. J., & Isaak, D. J. (2017) · Hydrological Processes · doi:10.1002/hyp.11139 · Cited by 80
“It is assumed that coefficients and intercepts in Equations 2 and 3 followed a Gaussian distribution (Gelman, 2006; Kwon, Brown, & Lall, 2008):”
BackgroundZhou, G., & Zhao, H. (2021) · PLoS Genetics · doi:10.1371/journal.pgen.1009697 · Cited by 84
Regulation of Speech and Media Coverage of Corruption: An Empirical Analysis of the Mexican Press
Stanig, P. (2014) · American Journal of Political Science · doi:10.1111/ajps.12110 · Cited by 93
Bayesian Data-Model Fit Assessment for Structural Equation Modeling
Levy, R. (2011) · Structural Equation Modeling A Multidisciplinary Journal · doi:10.1080/10705511.2011.607723 · Cited by 93
“Alternative choices for prior distributions have been proposed (e.g., see Gelman, 2006, for discussions and recommendations on alternative choices for priors on variance parameters in a variety of scenarios).”
BackgroundWhy Do Cuckolded Males Provide Paternal Care?
Griffin, A. S., Alonzo, S. H., & Cornwallis, C. K. (2013) · PLoS Biology · doi:10.1371/journal.pbio.1001520 · Cited by 77
An effect size measure and Bayesian analysis of single-case designs
Swaminathan, H., Rogers, H. J., & Horner, R. H. (2014) · Journal of School Psychology · doi:10.1016/j.jsp.2013.12.002 · Cited by 85
Hierarchical statistical modeling of xylem vulnerability to cavitation
Ogle, K., Barber, J. J., Willson, C. J., & Thompson, B. (2009) · New Phytologist · doi:10.1111/j.1469-8137.2008.02760.x · Cited by 75
A selection model for accounting for publication bias in a full network meta‐analysis
Mavridis, D., Welton, N. J., Sutton, A., & Salanti, G. (2014) · Statistics in Medicine · doi:10.1002/sim.6321 · Cited by 76
Armal, S., Devineni, N., & Khanbilvardi, R. M. (2017) · Journal of Climate · doi:10.1175/jcli-d-17-0106.1 · Cited by 79
“We assumed a uniform prior distribution for the variance terms and uninformative normal priors for the coefficients of the second level (Gelman 2006).”
MethodJin, Y., Russo, S. E., & Yu, M. (2017) · Journal of Ecology · doi:10.1111/1365-2745.12911 · Cited by 84
“Weakly informative R- side and parameterexpanded G- side priors improve MCMC convergence and chain mixing, especially for parameter- rich models, while allowing the data to dominate inference (Bolker, 2013; Gelman, 2006; McCarthy, 2013).”
BackgroundDownes, M., Gurrin, L. C., English, D. R., Pirkis, J., Currier, D., Spittal, M. J., & Carlin, J. B. (2018) · American Journal of Epidemiology · doi:10.1093/aje/kwy070 · Cited by 91
Bayesian latent variable modelling of multivariate spatio-temporal variation in cancer mortality
Tzala, E., & Best, N. (2008) · Statistical Methods in Medical Research · doi:10.1177/0962280207081243 · Cited by 85
Get Over It! A Multilevel Threshold Autoregressive Model for State-Dependent Affect Regulation
de Haan-Rietdijk, S., Gottman, J. M., Bergeman, C. S., & Hamaker, E. L. (2014) · Psychometrika · doi:10.1007/s11336-014-9417-x · Cited by 82
“…be noted that the inverse gamma and Wishart priors for variance parameters have been criticized because they may actually bias (co)variance estimates (Gelman, 2006), but they are oft-used and recommended proper priors in OpenBUGS for which a better alternative is not yet available to our knowledge.”
BackgroundGonadectomy effects on the risk of immune disorders in the dog: a retrospective study
Sundburg, C. R., Belanger, J. M., Bannasch, D. L., Famula, T. R., & Oberbauer, A. M. (2016) · BMC Veterinary Research · doi:10.1186/s12917-016-0911-5 · Cited by 83
“The next step in this hierarchical model, intending to address the challenge of evaluating uncommon disorders, we considered the weakly informative prior of a half-Cauchy for σ [43], that is σ ~ Cauchy(0,25).”
BackgroundHierarchical models for relational event sequences
DuBois, C., Butts, C. T., McFarland, D. A., & Smyth, P. (2013) · Journal of Mathematical Psychology · doi:10.1016/j.jmp.2013.04.001 · Cited by 78
“2 Because the inverse gamma distribution is a conjugate prior to a Gaussian distribution having a known location parameter, the posterior distribution of σ can be written in closed form [39]:”
BackgroundGeneralized linear mixed model with a penalized Gaussian mixture as a random effects distribution
Komárek, A., & Lesaffre, E. M. E. H. (2007) · Computational Statistics & Data Analysis · doi:10.1016/j.csda.2007.10.024 · Cited by 77
“(18) However, other commonly used priors for components of the inverse variance (see Gelman, 2006) are possible as well.”
Socioeconomic inequalities in cause-specific mortality in 15 European cities
Marí-Dell’Olmo, M., Gotsens, M., Palència, L., Burström, B., Corman, D., Costa, G., Deboosere, P., Díez, È., Domínguez‐Berjón, M. F., Dzúrová, D., et al. (2015) · Journal of Epidemiology & Community Health · doi:10.1136/jech-2014-204312 · Cited by 91
Bayesian latent time joint mixed effect models for multicohort longitudinal data
Li, D., Iddi, S., Thompson, W. K., Donohue, M. C., & Initiative, F. T. A. D. N. (2019) · Statistical Methods in Medical Research · doi:10.1177/0962280217737566 · Cited by 75
Bayesian Spatial Survival Models for Political Event Processes
Darmofal, D. (2008) · American Journal of Political Science · doi:10.1111/j.1540-5907.2008.00368.x · Cited by 85
Computational techniques for spatial logistic regression with large data sets
Paciorek, C. J. (2006) · Computational Statistics & Data Analysis · doi:10.1016/j.csda.2006.11.008 · Cited by 88
“Note that this prior strongly penalizes very large values of τ 2 , but avoids the sharp peak near zero of priors of the form G ( (cid:15), (cid:15) ) with (cid:15) ≈ 0 (Gelman 2006).”
Soanes, K., Vesk, P. A., & van der Ree, R. (2015) · Wildlife Research · doi:10.1071/wr14067 · Cited by 74
Fear of self, doubt and obsessive compulsive symptoms
Nikodijevic, A., Moulding, R., Anglim, J., Aardema, F., & Nedeljkovic, M. (2015) · Journal of Behavior Therapy and Experimental Psychiatry · doi:10.1016/j.jbtep.2015.02.005 · Cited by 82
Honarmandi, P., & Arróyave, R. (2020) · Integrating materials and manufacturing innovation · doi:10.1007/s40192-020-00168-2 · Cited by 87
“For example, informative inverse gamma prior distributions also need to have hyper-parameters greater than 1 [75].”
BackgroundPrior Knowledge Elicitation: The Past, Present, and Future
Mikkola, P., Martin, O. A., Chandramouli, S., Hartmann, M., Abril-Pla, O., Thomas, O., Pesonen, H., Corander, J., Vehtari, A., Kaski, S., et al. (2023) · Bayesian Analysis · doi:10.1214/23-ba1381 · Cited by 165
“However, we repeat the general observation that flat priors (such as N(0, 106)) sometimes used by practitioners should be avoided, due to problems in posterior inference (Carlin, 2000; van Dongen, 2006; Gelman, 2006; Gelman et al., 2013, 2017; Gelman and Yao, 2020; Smid and Winter, 2020).”
BackgroundInfluentialA Bayesian Approach to Analyzing Ecotoxicological Data
Billoir, É., Delignette‐Muller, M. L., Péry, A. R. R., & Charles, S. (2008) · Environmental Science & Technology · doi:10.1021/es801418x · Cited by 74
Matonis, M., Walters, M. B., & Millington, J. D. A. (2011) · Forest Ecology and Management · doi:10.1016/j.foreco.2011.03.034 · Cited by 88
Comparison of confidence interval methods for an intra-class correlation coefficient (ICC)
Ionan, A. C., Polley, M. C., McShane, L. M., & Dobbin, K. K. (2014) · BMC Medical Research Methodology · doi:10.1186/1471-2288-14-121 · Cited by 84
Age and race differences in the trajectories of self-esteem.
Shaw, B. A., Liang, J., & Krause, N. (2010) · Psychology and Aging · doi:10.1037/a0018242 · Cited by 72
“First, although there is a consensus for weighting data in generating descriptive statistics for a given target population, there is no such agreement in multivariate analyses (Gelman, 2006).”
BackgroundRuncie, D. E., & Mukherjee, S. (2013) · Genetics · doi:10.1534/genetics.113.151217 · Cited by 76
“For example, de los Campos and Gianola (2007) fit an unconstrained R, although they used an informative inverse Wishart prior (Gelman 2006) and only consider five traits.”
MethodAn Epidemic Model for News Spreading on Twitter
Abdullah, S., & Wu, X. (2011) · IEEE International Conference on Tools with Artificial Intelligence (ICTAI) · doi:10.1109/ictai.2011.33 · Cited by 93
Bayesian estimation in random effects meta‐analysis using a non‐informative prior
Bodnar, O., Link, A., Arendacká, B., Possolo, A., & Elster, C. (2017) · Statistics in Medicine · doi:10.1002/sim.7156 · Cited by 73
Webb, W. H., Brunton, D. H., Aguirre, J. D., Thomas, D. B., Vâlcu, M., & Dale, J. (2016) · Frontiers in Ecology and Evolution · doi:10.3389/fevo.2016.00022 · Cited by 82
“For the variance components we used priors conforming to a scaled non-central F-distribution (Gelman, 2006) with the location parameter equal to zero.”
MethodIncorporating Nonergodic Path Effects into the NGA‐West2 Ground‐Motion Prediction Equations
Kuehn, N. M., Abrahamson, N., & Walling, M. (2019) · Bulletin of the Seismological Society of America · doi:10.1785/0120180260 · Cited by 73
A data-based predictive model for spatiotemporal variability in stream water quality
Guo, D., Lintern, A. M., Webb, J. A., Ryu, D., Bende‐Michl, U., Liu, S., & Western, A. W. (2020) · Hydrology and earth system sciences · doi:10.5194/hess-24-827-2020 · Cited by 76
“The constituent standard deviation (σ ) was assumed to be drawn from a minimally informative half-normal prior distribution of N (0, 10) truncated to only positive values (Gelman, 2006; Stan Development Team, 2018).”
BackgroundMethodKass, R. E., & Natarajan, R. (2006) · Bayesian Analysis · doi:10.1214/06-ba117b · Cited by 72
“guidance as to when posteriors based on priors for D, including the uniform prior, the default conjugate prior, or other interesting choices such as that recommended by Gelman (2005), are likely to have good frequentist operating characteristics.”
BackgroundAssessing Multivariate Constraints to Evolution across Ten Long-Term Avian Studies
Teplitsky, C., Tarka, M., Møller, A. P., et al. (2014) · PLoS ONE · doi:10.1371/journal.pone.0090444 · Cited by 73
Ammad-ud-din, M., Khan, S. A., Wennerberg, K., & Aittokallio, T. (2017) · Bioinformatics · doi:10.1093/bioinformatics/btx266 · Cited by 73
Animal Models and Integrated Nested Laplace Approximations
Holand, A. M., Steinsland, I., Martino, S., & Jensen, H. (2013) · G3 Genes Genomes Genetics · doi:10.1534/g3.113.006700 · Cited by 74
“Priors for variances are discussed in Gelman (2006), and this topic should be further investigated, but it is outside the scope of this work.”
Schmitz, J. M., Green, C. E., Hasan, K. M., Vincent, J. N., Suchting, R., Weaver, M. F., Moeller, F. G., Narayana, P. A., Cunningham, K. A., Dineley, K. T., et al. (2017) · Addiction · doi:10.1111/add.13868 · Cited by 73
MCMC for Imbalanced Categorical Data
Johndrow, J. E., Smith, A. D., Pillai, N. S., & Dunson, D. B. (2018) · Journal of the American Statistical Association · doi:10.1080/01621459.2018.1505626 · Cited by 92
“…across the serving sites via a hierarchical model: yi | ni, pi ∼ Binomial ( ni, g−1(θi) ) , θ ∼ No(b1N,BIN ) (1) b ∼ No(b0,B0), B ∼ π(B), where π(B) is a half Cauchy prior on √ B, recommended by Gelman (2006) and Polson and Scott (2012), 1N denotes an Nvector of 1’s, and IN is the N × N identity.”
ResultDawson, A., Paciorek, C. J., McLACHLAN, J. S., Goring, S. J., Williams, J. W., & Jackson, S. T. (2016) · Quaternary Science Reviews · doi:10.1016/j.quascirev.2016.01.012 · Cited by 77
Evolution of Genetic Variance during Adaptive Radiation
Walter, G. M., Aguirre, J. D., Blows, M. W., & Ortíz-Barrientos, D. (2018) · The American Naturalist · doi:10.1086/696123 · Cited by 77
Hydrological drivers of record‐setting water level rise on Earth's largest lake system
Gronewold, A. D., Bruxer, J., Durnford, D. A., Smith, J. P., Clites, A. H., Seglenieks, F. R., Qian, S. S., Hunter, T. S., & Fortin, V. (2016) · Water Resources Research · doi:10.1002/2015wr018209 · Cited by 84
GBAGC: A General Bayesian Framework for Attributed Graph Clustering
Xu, Z., Ke, Y., Wang, Y., Cheng, H., & Cheng, J. S.-C. (2014) · ACM Transactions on Knowledge Discovery from Data · doi:10.1145/2629616 · Cited by 68
“…than 1 so that the prior p(fkl) provides as little information as possible about the true values of parameters fkl and let the data speak for themselves [Gelman 2006]. updating rules for weighted attributed graphs: N . ktm = .tm + ß ikd(Yit,atm) , i=1 N µ kk = µ+ ß ikß jkXij, i, j=1 i<j N . kk…”
BackgroundLocally Adaptive Spatial Smoothing Using Conditional Auto-Regressive Models
Lee, D., & Mitchell, R. (2013) · Journal of the Royal Statistical Society Series C (Applied Statistics) · doi:10.1111/rssc.12009 · Cited by 85
Brassil, C. E., & Couch, B. A. (2019) · International Journal of STEM Education · doi:10.1186/s40594-019-0169-0 · Cited by 91
“Half-Cauchy priors (Gelman, 2006) were used as weak-information priors (σ = 2.5) for the dispersal parameter of the hierarchical probability values (i.e., the distributions from which the question-level values were drawn).”
MethodInfluentialDiminishing benefits of urban living for children and adolescents' growth and development
NCD Risk Factor Collaboration (NCD-RisC), Mishra, A., Zhou, B., Rodriguez-Martinez, A., Bixby, H., Singleton, R. K., Carrillo-Larco, R. M., Sheffer, K. E., Paciorek, C. J., Bennett, J. E., et al. (2023) · Nature · doi:10.1038/s41586-023-05772-8 · Cited by 70
Achana, F. A., Cooper, N. J., Bujkiewicz, S., Hubbard, S., Kendrick, D., Jones, D. R., & Sutton, A. J. (2014) · BMC Medical Research Methodology · doi:10.1186/1471-2288-14-92 · Cited by 81
Quantifying demographic uncertainty: Bayesian methods for integral projection models
Elderd, B. D., & Miller, T. E. X. (2015) · Ecological Monographs · doi:10.1890/15-1526.1 · Cited by 80
LARGE POPULATION SIZE PREDICTS THE DISTRIBUTION OF ASEXUALITY IN SCALE INSECTS
Ross, L., Hardy, N. B., Okusu, A., & Normark, B. B. (2012) · Evolution · doi:10.1111/j.1558-5646.2012.01784.x · Cited by 73
“We used parameter expanded priors for the variances of the random effects for the taxonomic model, which are relatively uniform for the standard deviations (Gelman 2006; Hadfield 2010b).”
MethodBayesian Covariance Selection in Generalized Linear Mixed Models
Cai, B., & Dunson, D. B. (2005) · Biometrics · doi:10.1111/j.1541-0420.2005.00499.x · Cited by 69
“Preliminary work has focused on simple normal variance component and independent normal linear random effects models (Pauler, Wakefield, and Kass, 1999; Sun, Tsutakawa, and He, 2001; Gelman, 2006).”
MethodOccupant-level injury severity analyses for taxis in Hong Kong: A Bayesian space-time logistic model
Meng, F., Xu, P., Wong, S. C., et al. (2017) · Accident Analysis & Prevention · doi:10.1016/j.aap.2017.08.010 · Cited by 70
“However, the inverse-Gamma distribution has been proven sensitive to 221 the parameter if and are close to zero (Gelman 2006, Lee 2011).”
BackgroundZitzmann, S., Lüdtke, O., Robitzsch, A., & Hecht, M. (2020) · Structural Equation Modeling A Multidisciplinary Journal · doi:10.1080/10705511.2020.1752216 · Cited by 73
“…statistical literature, the question of how prior distributions must be specified has been discussed in the past, their use has been evaluated in simulation studies, and recommendations have been made (e.g., Baldwin & Fellingham, 2013; Browne & Draper, 2006; Gelman, 2006; Natarajan & Kass, 2000).”
BackgroundA comparison of spatial smoothing methods for small area estimation with sampling weights
Mercer, L. D., Wakefield, J. C., Chen, C., & Lumley, T. (2013) · Spatial Statistics · doi:10.1016/j.spasta.2013.12.001 · Cited by 80
“Lambert (2006) prefers uniform priors on the standard deviation, σ, which is further supported by Gelman (2006) who states: “In fitting hierarchical models, we recommend starting with a noninformative uniform prior density on standard deviation parameters σ”.”
BackgroundSantana, P., da Costa, C. H., Marí-Dell’Olmo, M., Gotsens, M., & Borrell, C. (2015) · International Journal for Equity in Health · doi:10.1186/s12939-015-0182-y · Cited by 87
An Adaptive Spatiotemporal Smoothing Model for Estimating Trends and Step Changes in Disease Risk
Rushworth, A., Lee, D., & Sarran, C. E. (2016) · Journal of the Royal Statistical Society Series C (Applied Statistics) · doi:10.1111/rssc.12155 · Cited by 69
“Our choice of inverse gamma priors here is due to their conjugacy, but alternatives could be explored such as the half-Cauchy distribution as suggested by Gelman (2006).”
BackgroundPredicting Species Distributions from Samples Collected along Roadsides
McCarthy, K. P., Fletcher, R. J., Rota, C. T., & Hutto, R. L. (2011) · Conservation Biology · doi:10.1111/j.1523-1739.2011.01754.x · Cited by 81
Chou, W.-C., & Lin, Z. (2020) · Environment International · doi:10.1016/j.envint.2020.105581 · Cited by 68
A Bayesian Kriging model applied for spatial downscaling of daily rainfall from GCMs
Lima, C., Kwon, H., & Kim, Y. (2021) · Journal of Hydrology · doi:10.1016/j.jhydrol.2021.126095 · Cited by 72
Does uncertainty affect real activity? Evidence from state-level data
Mumtaz, H. (2018) · Economics Letters · doi:10.1016/j.econlet.2018.03.026 · Cited by 68
Bayesian spatiotemporal mapping of relative dengue disease risk in Bandung, Indonesia
Jaya, I. G. N. M., & Folmer, H. (2019) · Journal of Geographical Systems · doi:10.1007/s10109-019-00311-4 · Cited by 74
“2018; Gelman 2006).”
BackgroundProducing Standardized Country-Level Immunization Delivery Unit Cost Estimates
Portnoy, A., Vaughan, K., Clarke‐Deelder, E., Suharlim, C., Resch, S., Brenzel, L., & Menzies, N. A. (2020) · PharmacoEconomics · doi:10.1007/s40273-020-00930-6 · Cited by 62
Virulence mismatches in index hosts shape the outcomes of cross-species transmission
Mollentze, N., Streicker, D. G., Murcia, P. R., Hampson, K., & Biek, R. (2020) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.2006778117 · Cited by 61
Influence of artificial food provisioning from fisheries on killer whale reproductive output
Tixier, P., Authier, M., Gasco, N., & Guinet, C. (2014) · Animal Conservation · doi:10.1111/acv.12161 · Cited by 75
“If the data are informative, these priors have a negligible influence (Gelman, 2006; Gelman et al., 2008).”
BackgroundYang, Y. (2006) · Sociological Methodology · doi:10.1111/j.1467-9531.2006.00174.x · Cited by 73
“Gelman (2006) showed the difficulty of the inverse-gamma (ε, ε) prior distribution when the limit of ε → 0, in which case it yields an improper posterior density.”
BackgroundPéron, C., Authier, M., Barbraud, C., Delord, K., Besson, D., & Weimerskirch, H. (2010) · Global Change Biology · doi:10.1111/j.1365-2486.2010.02169.x · Cited by 71
“We used weakly informative priors (Gelman, 2006) and inferences were drawn for posterior distributions based on 1 000 000 Markov Chain Monte Carlo simulations with every 500th sample retained to reduce autocorrelation (see Inference details and WinBUGS codes in supplementary information).”
MethodA Bayesian Antedependence Model for Whole Genome Prediction
Yang, W., & Tempelman, R. J. (2012) · Genetics · doi:10.1534/genetics.111.131540 · Cited by 66
“Finally, we specify noninformative priors s 2 e x 2 2 ð 2 1 ; 0 Þ and s 2 u x 2 2 ð 2 1 ; 0 Þ , which are congruent with specifying uniform priors on s e and s u , respectively, and in line with recommendations for variance components by Gelman (2006).”
Method
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.