pith. sign in

arxiv: 2606.28015 · v1 · pith:REVIZMACnew · submitted 2026-06-26 · 📊 stat.ME

Bayesian Simultaneous Credible Bands for Polynomial Regression

Fei Yang , Yang Han , Wei Liu , Ian Hall This is my paper

Pith reviewed 2026-06-29 02:55 UTC · model grok-4.3

classification 📊 stat.ME
keywords bayesian credible bandssimultaneous inferencepolynomial regressionfrequentist coveragedose-response studiescredible intervals
0
0 comments X

The pith

Bayesian simultaneous credible bands for polynomial regression match frequentist coverage asymptotically while delivering exact posterior coverage.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper constructs a unified framework for Bayesian simultaneous credible bands on the regression curve of a univariate polynomial model over a finite interval. It supplies a simulation-based method to compute the critical constant, usable either analytically or from posterior draws, and accommodates varying amounts of prior information. The central result shows that under mild regularity conditions the Bayesian band is asymptotically equivalent to the frequentist simultaneous confidence band, so it inherits the nominal frequentist coverage for a broad class of priors. Simulations verify that the band achieves its exact posterior simultaneous coverage probability in finite samples. An application to dose-response data shows how the bands can locate the minimum effective dose.

Core claim

Under mild regularity conditions, the BSCB is asymptotically equivalent to the FSCB, thereby attaining the nominal frequentist coverage for a broad class of priors. The BSCB attains the exact posterior simultaneous coverage probability across various scenarios.

What carries the argument

The unified framework for BSCB construction that uses a simulation-based procedure to determine the critical constant, implementable analytically or via posterior sampling.

If this is right

  • The bands allow prior information to be used in simultaneous inference for polynomial dose-response curves.
  • The same critical-constant procedure works for both analytic and MCMC-based posterior sampling.
  • The bands can locate the minimum effective dose in Phase II trials while retaining frequentist coverage properties.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The equivalence argument may extend to other parametric families if similar regularity conditions can be verified.
  • Informative priors could produce narrower bands than frequentist versions while preserving the coverage guarantees.

Load-bearing premise

The framework applies only to univariate polynomial models on a finite covariate interval and the asymptotic equivalence rests on unspecified mild regularity conditions.

What would settle it

A concrete simulation in which the BSCB fails to achieve its nominal posterior simultaneous coverage probability under the paper's regularity conditions would falsify the exact-coverage claim.

Figures

Figures reproduced from arXiv: 2606.28015 by Fei Yang, Ian Hall, Wei Liu, Yang Han.

Figure 1
Figure 1. Figure 1: Comparison of three types of BSCB with BPCB-I-J for the quadratic model [PITH_FULL_IMAGE:figures/full_fig_p021_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Comparison of three types of BSCB with BPCB-I-J for the cubic model [PITH_FULL_IMAGE:figures/full_fig_p022_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: The 95% BSCB-H-N, BSCB-C-U, BSCB-I-J, FSCB-E, and BPCB-I-J for the [PITH_FULL_IMAGE:figures/full_fig_p028_3.png] view at source ↗
read the original abstract

Quantifying efficacy uncertainty across the entire dose range is crucial in dose-response studies. Although the frequentist simultaneous confidence band (FSCB) is widely used for this purpose, it does not readily incorporate prior knowledge. The Bayesian simultaneous credible band (BSCB) offers a natural alternative, yet practical methods for constructing BSCBs remain scarce in the literature. In this paper, we propose a unified framework for constructing a BSCB for the regression curve in a univariate polynomial model over a finite covariate interval. An efficient simulation-based procedure is developed to determine the critical constant of a BSCB. The framework accommodates inference under different levels of prior information and can be implemented either analytically or via posterior sampling methods. Notably, we prove that under mild regularity conditions, the BSCB is asymptotically equivalent to the FSCB, thereby attaining the nominal frequentist coverage for a broad class of priors. Simulation studies confirm that the BSCB attains the exact posterior simultaneous coverage probability across various scenarios. An application to a dose-response study illustrates its importance in identifying the minimum effective dose in Phase II clinical trials. Software implementation of the proposed methods is available in an accompanying R package.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The manuscript proposes a unified framework for Bayesian simultaneous credible bands (BSCB) in univariate polynomial regression over a finite covariate interval. It develops a simulation-based procedure to compute the critical constant, supports analytic or posterior-sampling implementations under varying prior information, proves that the BSCB is asymptotically equivalent to the frequentist simultaneous confidence band (FSCB) under mild regularity conditions for a broad class of priors (thereby inheriting nominal frequentist coverage), validates exact posterior simultaneous coverage via simulations, and demonstrates the method on a dose-response study for minimum effective dose identification, accompanied by an R package.

Significance. If the asymptotic equivalence result holds under the stated conditions, the work supplies a practical bridge between Bayesian and frequentist simultaneous inference for polynomial models, enabling prior incorporation while preserving frequentist coverage properties asymptotically. This is relevant for dose-response applications in clinical trials. Credit is due for the accompanying R package (enhancing reproducibility) and the simulation studies confirming posterior coverage across scenarios.

major comments (2)
  1. [Theoretical development of asymptotic equivalence] The central asymptotic equivalence result (stated in the abstract and presumably proved in the theoretical section) rests on 'mild regularity conditions' that are neither explicitly enumerated nor verified for the polynomial regression setting or the priors employed. Without a precise statement (e.g., moment or tail conditions on the prior, smoothness or design requirements), the scope of the claimed equivalence for a 'broad class of priors' cannot be assessed and the result remains formally unverified for the manuscript's framework.
  2. [Simulation studies section] The simulation studies are invoked to confirm exact posterior simultaneous coverage, yet the manuscript provides no details on data exclusion rules, error analysis, or sensitivity to polynomial degree selection. These omissions undermine evaluation of whether the reported coverage holds robustly across the scenarios considered.
minor comments (2)
  1. [Abstract] The abstract refers to 'mild regularity conditions' without any characterization; a one-sentence indication of their nature would improve readability.
  2. [Methodological framework] Notation for the critical constant and the polynomial basis should be checked for consistency between the analytic and sampling-based implementations.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the detailed and constructive report. We address each major comment below, indicating where revisions will be made to strengthen the manuscript.

read point-by-point responses

  1. Referee: [Theoretical development of asymptotic equivalence] The central asymptotic equivalence result (stated in the abstract and presumably proved in the theoretical section) rests on 'mild regularity conditions' that are neither explicitly enumerated nor verified for the polynomial regression setting or the priors employed. Without a precise statement (e.g., moment or tail conditions on the prior, smoothness or design requirements), the scope of the claimed equivalence for a 'broad class of priors' cannot be assessed and the result remains formally unverified for the manuscript's framework.

    Authors: We agree that the regularity conditions require a more explicit statement to allow readers to assess the scope of the asymptotic equivalence result. In the revised manuscript we will add a dedicated subsection that enumerates the precise conditions (moment and tail conditions on the prior, design density requirements, and smoothness assumptions on the regression function) and verify that they hold for the polynomial model and the class of priors considered. revision: yes

  2. Referee: [Simulation studies section] The simulation studies are invoked to confirm exact posterior simultaneous coverage, yet the manuscript provides no details on data exclusion rules, error analysis, or sensitivity to polynomial degree selection. These omissions undermine evaluation of whether the reported coverage holds robustly across the scenarios considered.

    Authors: We acknowledge that additional implementation details are needed for reproducibility and robustness assessment. In the revision we will expand the simulation section to include explicit data exclusion criteria (if any), Monte Carlo error analysis, and a sensitivity study with respect to polynomial degree, reporting coverage results across a range of degrees. revision: yes

Circularity Check

0 steps flagged

No significant circularity detected

full rationale

The paper's central claim is a mathematical proof that BSCB is asymptotically equivalent to FSCB under mild regularity conditions for a broad class of priors, presented as an independent derivation rather than a self-definition, fitted input renamed as prediction, or load-bearing self-citation. No equations or steps in the abstract reduce the result to its own inputs by construction. The framework for constructing BSCB via simulation or posterior sampling is described as a new procedure, and simulation studies are used for confirmation rather than defining the result. This is the expected honest non-finding for a paper whose derivation chain rests on external proof rather than tautology.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review provides no information on specific free parameters, axioms, or invented entities; full manuscript required to audit these elements.

pith-pipeline@v0.9.1-grok · 5727 in / 1186 out tokens · 55890 ms · 2026-06-29T02:55:01.398763+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

300 extracted references · 21 canonical work pages · 1 internal anchor

  1. [1]

    2021 , publisher=

    An introduction to Bayesian inference, methods and computation , author=. 2021 , publisher=

    2021

  2. [2]

    Journal of the Royal Statistical Society: Series B (Methodological) , volume=

    The use of the concept of a future observation in goodness-of-fit problems , author=. Journal of the Royal Statistical Society: Series B (Methodological) , volume=. 1967 , publisher=

    1967

  3. [3]

    The Annals of Statistics , pages=

    Bayesianly justifiable and relevant frequency calculations for the applied statistician , author=. The Annals of Statistics , pages=. 1984 , publisher=

    1984

  4. [4]

    2014 , publisher=

    Bayesian essentials with R , author=. 2014 , publisher=

    2014

  5. [5]

    Bayesian inference and decision techniques , year=

    On assessing prior distributions and Bayesian regression analysis with g-prior distributions , author=. Bayesian inference and decision techniques , year=

  6. [6]

    Journal of the american statistical association , volume=

    A reference Bayesian test for nested hypotheses and its relationship to the Schwarz criterion , author=. Journal of the american statistical association , volume=. 1995 , publisher=

    1995

  7. [7]

    Journal of the American Statistical Association , volume=

    Interactive elicitation of opinion for a normal linear model , author=. Journal of the American Statistical Association , volume=. 1980 , publisher=

    1980

  8. [8]

    2008 , publisher =

    Bayesian Methods for Data Analysis , author =. 2008 , publisher =

    2008

  9. [9]

    Bayesian Analysis , number =

    Andrew Gelman , title =. Bayesian Analysis , number =. 2006 , doi =

    2006

  10. [10]

    1973 , publisher =

    Bayesian Inference in Statistical Analysis , author =. 1973 , publisher =. doi:10.1002/9781118033197 , note =

    work page doi:10.1002/9781118033197 1973

  11. [11]

    2009 , publisher=

    Computation of multivariate normal and t probabilities , author=. 2009 , publisher=

    2009

  12. [12]

    Statistics in medicine , volume=

    Bayesian hierarchical EMAX model for dose-response in early phase efficacy clinical trials , author=. Statistics in medicine , volume=. 2019 , publisher=

    2019

  13. [13]

    Biometrics , volume=

    Robust meta-analytic-predictive priors in clinical trials with historical control information , author=. Biometrics , volume=. 2014 , publisher=

    2014

  14. [14]

    Statistical science , volume=

    Inference from iterative simulation using multiple sequences , author=. Statistical science , volume=. 1992 , publisher=

    1992

  15. [15]

    Journal of the Royal Statistical Society Series C: Applied Statistics , volume=

    Bayesian modelling of effective and functional brain connectivity using hierarchical vector autoregressions , author=. Journal of the Royal Statistical Society Series C: Applied Statistics , volume=. 2024 , publisher=

    2024

  16. [16]

    Journal of Computational and Graphical Statistics , volume=

    Simultaneous posterior probability statements from Monte Carlo output , author=. Journal of Computational and Graphical Statistics , volume=. 2004 , publisher=

    2004

  17. [17]

    Journal of Nonparametric Statistics , volume=

    Simultaneous bootstrap confidence bands in nonparametric regression , author=. Journal of Nonparametric Statistics , volume=. 1998 , publisher=

    1998

  18. [18]

    Journal of Computational and Graphical Statistics , volume=

    Fast hamiltonian monte carlo using gpu computing , author=. Journal of Computational and Graphical Statistics , volume=. 2016 , publisher=

    2016

  19. [19]

    2025 , note =

    RStan: the. 2025 , note =

    2025

  20. [20]

    Biometrics , volume=

    Horseshoe-based Bayesian nonparametric estimation of effective population size trajectories , author=. Biometrics , volume=. 2020 , publisher=

    2020

  21. [21]

    1961 , publisher =

    Jeffreys, Harold , title =. 1961 , publisher =

    1961

  22. [22]

    Biometrika , volume=

    Efficient parametrisations for normal linear mixed models , author=. Biometrika , volume=. 1995 , publisher=

    1995

  23. [23]

    2009 , publisher =

    Computation of Multivariate Normal and t Probabilities , author =. 2009 , publisher =

    2009

  24. [24]

    2013 , edition =

    Bayesian Data Analysis , author =. 2013 , edition =. doi:10.1201/b16018 , url =

    work page doi:10.1201/b16018 2013

  25. [25]

    Journal of the American statistical association , volume=

    Rectangular confidence regions for the means of multivariate normal distributions , author=. Journal of the American statistical association , volume=. 1967 , publisher=

    1967

  26. [26]

    Encyclopedia of statistical sciences , year=

    Bonferroni Inequalities and Intervals , author=. Encyclopedia of statistical sciences , year=

  27. [27]

    , author=

    The No-U-Turn sampler: adaptively setting path lengths in Hamiltonian Monte Carlo. , author=. J. Mach. Learn. Res. , volume=

  28. [28]

    Proceedings of the Royal Society of London

    An invariant form for the prior probability in estimation problems , author=. Proceedings of the Royal Society of London. Series A. Mathematical and Physical Sciences , volume=. 1946 , publisher=

    1946

  29. [29]

    1980 , publisher=

    Approximation Theorems of Mathematical Statistics , author=. 1980 , publisher=

    1980

  30. [30]

    2003 , publisher=

    Bayesian nonparametrics , author=. 2003 , publisher=

    2003

  31. [31]

    Handbook of econometrics , volume=

    Large sample estimation and hypothesis testing , author=. Handbook of econometrics , volume=. 1994 , publisher=

    1994

  32. [32]

    Vaart, A. W. van der , year=. Asymptotic Statistics , publisher=

  33. [33]

    2021 , publisher =

    Regression: Models, Methods and Applications , author =. 2021 , publisher =

    2021

  34. [34]

    American journal of epidemiology , volume=

    Flexible meta-regression functions for modeling aggregate dose-response data, with an application to alcohol and mortality , author=. American journal of epidemiology , volume=. 2004 , publisher=

    2004

  35. [35]

    Clinical Trials , volume=

    Strategic inclusion of regions in multiregional clinical trials , author=. Clinical Trials , volume=. 2019 , publisher=

    2019

  36. [36]

    1994 , howpublished =

    ICH , title =. 1994 , howpublished =

    1994

  37. [37]

    1985 , publisher=

    Statistical decision theory and Bayesian analysis , author=. 1985 , publisher=

    1985

  38. [38]

    2024 , publisher=

    Objective Bayesian Inference , author=. 2024 , publisher=

    2024

  39. [39]

    Journal of the American Statistical Association , volume=

    Minimum area confidence set optimality for confidence bands in simple linear regression , author=. Journal of the American Statistical Association , volume=. 2007 , publisher=

    2007

  40. [40]

    Biometrics , volume=

    Combining multiple comparisons and modeling techniques in dose-response studies , author=. Biometrics , volume=. 2005 , publisher=

    2005

  41. [41]

    Crop science , volume=

    Response of Rice to Solar Radiation and Temperature Estimated from International Yield Trials 1 , author=. Crop science , volume=. 1984 , publisher=

    1984

  42. [42]

    Statistical methods for the analysis of biomedical data, second edition , year =

    Woolson, Robert F and Clarke, William R , address =. Statistical methods for the analysis of biomedical data, second edition , year =

  43. [43]

    All of Nonparametric Statistics , isbn =

    Wasserman, Larry , year =. All of Nonparametric Statistics , isbn =. Pittsburgh , doi =

  44. [44]

    2025 , eprint=

    Quasi-Bayesian Local Projections: Simultaneous Inference and Extension to the Instrumental Variable Method , author=. 2025 , eprint=

    2025

  45. [45]

    Minnema, L. A. and Giezen, T. J. and Gardarsdottir, H. and Leufkens, H. G. M. and Egberts, T. C. G. and Mantel-Teeuwisse, A. K. , title =. British Journal of Clinical Pharmacology , year =

  46. [46]

    Sacks and Hala H

    Leonard V. Sacks and Hala H. Shamsuddin and Yuliya I. Yasinskaya and Khaled Bouri and Michael L. Lanthier and Rachel E. Sherman , title =. JAMA , year =

  47. [47]

    Clinical Trials , volume=

    Dose-response characterization: A key to success in drug development , author=. Clinical Trials , volume=. 2025 , publisher=

    2025

  48. [48]

    Metrika , volume=

    One-sided hyperbolic simultaneous confidence bands for multiple and polynomial regression models , author=. Metrika , volume=. 2017 , publisher=

    2017

  49. [49]

    2005 , publisher=

    Analyzing environmental data , author=. 2005 , publisher=

    2005

  50. [50]

    Journal of Applied Econometrics , volume=

    Simultaneous confidence bands: Theory, implementation, and an application to SVARs , author=. Journal of Applied Econometrics , volume=. 2019 , publisher=

    2019

  51. [51]

    Neural Networks , volume=

    How to evaluate uncertainty estimates in machine learning for regression? , author=. Neural Networks , volume=. 2024 , publisher=

    2024

  52. [52]

    American Journal of Mathematics , volume=

    On the volume of tubes , author=. American Journal of Mathematics , volume=. 1939 , publisher=

    1939

  53. [53]

    American Journal of Mathematics , volume=

    Tubes and spheres in n-spaces, and a class of statistical problems , author=. American Journal of Mathematics , volume=. 1939 , publisher=

    1939

  54. [54]

    Scandinavian Journal of Statistics , volume=

    Simultaneous credible bands for latent Gaussian models , author=. Scandinavian Journal of Statistics , volume=. 2011 , publisher=

    2011

  55. [55]

    and Klauenberg, K

    Elster, C. and Klauenberg, K. and Walzel, M. and Wübbeler, G. and Harris, P. and Cox, M. and Matthews, C. and Smith, I. and Wright, L. and Allard, A. and Fischer, N. and Cowen, S. and Ellison, S. and Wilson, P. and Pennecchi, F. and Kok, G. and van der Veen, A. and Pendrill, L. , title =. 2015 , month =

    2015

  56. [56]

    Journal of Computational and Graphical Statistics , volume=

    Spatially adaptive Bayesian penalized splines with heteroscedastic errors , author=. Journal of Computational and Graphical Statistics , volume=. 2007 , publisher=

    2007

  57. [57]

    Statistical science , pages=

    Bayesian computation and stochastic systems , author=. Statistical science , pages=. 1995 , publisher=

    1995

  58. [58]

    Journal of the American Statistical Association , volume=

    Simultaneous confidence bands for penalized spline estimators , author=. Journal of the American Statistical Association , volume=. 2010 , publisher=

    2010

  59. [59]

    2011 , publisher=

    Bayesian inference in statistical analysis , author=. 2011 , publisher=

    2011

  60. [60]

    Australian & New Zealand Journal of Statistics , volume =

    Liu, Wei and Zhou, Sanyu and Bretz, Frank , title =. Australian & New Zealand Journal of Statistics , volume =. doi:https://doi.org/10.1111/anzs.12048 , url =. https://onlinelibrary.wiley.com/doi/pdf/10.1111/anzs.12048 , year =

    work page doi:10.1111/anzs.12048

  61. [61]

    Johnstone , journal =

    Søren Johansen and Iain M. Johnstone , journal =. Hotelling's Theorem on the Volume of Tubes: Some Illustrations in Simultaneous Inference and Data Analysis , urldate =

  62. [62]

    On Hotelling's Approach to Testing for a Nonlinear Parameter in Regression , urldate =

    Mark Knowles and David Siegmund , journal =. On Hotelling's Approach to Testing for a Nonlinear Parameter in Regression , urldate =

  63. [63]

    On Hotelling's Formula for the Volume of Tubes and Naiman's Inequality , urldate =

    Iain Johnstone and David Siegmund , journal =. On Hotelling's Formula for the Volume of Tubes and Naiman's Inequality , urldate =

  64. [64]

    Naiman , title =

    Daniel Q. Naiman , title =. The Annals of Statistics , pages =

  65. [65]

    Communications in Statistics - Theory and Methods , volume =

    Shan Lin , title =. Communications in Statistics - Theory and Methods , volume =. 2016 , publisher =. doi:10.1080/03610926.2013.833242 , URL =

    work page doi:10.1080/03610926.2013.833242 2016

  66. [66]

    Journal of Statistical Computation and Simulation , volume=

    Statistical inferences for linear regression models when the covariates have functional relationships: polynomial regression , author=. Journal of Statistical Computation and Simulation , volume=. 2008 , publisher=

    2008

  67. [67]

    Journal of Computational and Graphical Statistics , volume=

    Simulation-based simultaneous confidence bands in multiple linear regression with predictor variables constrained in intervals , author=. Journal of Computational and Graphical Statistics , volume=. 2005 , publisher=

    2005

  68. [68]

    Statistics in medicine , volume=

    Exact simultaneous confidence bands for a collection of univariate polynomials in regression analysis , author=. Statistics in medicine , volume=. 1999 , publisher=

    1999

  69. [69]

    The Annals of Statistics , pages=

    Simultaneous confidence bands for linear regression and smoothing , author=. The Annals of Statistics , pages=. 1994 , publisher=

    1994

  70. [70]

    Technometrics , volume=

    Uniform confidence bands for a quadratic model , author=. Technometrics , volume=. 1973 , publisher=

    1973

  71. [71]

    Biometrika , volume=

    An exact confidence band for one-dimensional polynomial regression , author=. Biometrika , volume=. 1984 , publisher=

    1984

  72. [72]

    Y., John Wiley & Sons , volume=

    The analysis of variance New York , author=. Y., John Wiley & Sons , volume=

  73. [73]

    Spurrier , title =

    John D. Spurrier , title =. Technometrics , volume =. 1993 , publisher =. doi:10.1080/00401706.1993.10485327 , URL =

    work page doi:10.1080/00401706.1993.10485327 1993

  74. [74]

    Wynn and P

    Henry P. Wynn and P. Bloomfield , journal =. Simultaneous Confidence Bands in Regression Analysis , urldate =

  75. [75]

    Journal of the American Statistical Association , volume =

    Esa Uusipaikka and , title =. Journal of the American Statistical Association , volume =. 1983 , publisher =. doi:10.1080/01621459.1983.10478023 , URL =

    work page doi:10.1080/01621459.1983.10478023 1983

  76. [76]

    Applications of the Theory of Error to the Interpretation of Trends , urldate =

    Holbrook Working and Harold Hotelling , journal =. Applications of the Theory of Error to the Interpretation of Trends , urldate =

  77. [77]

    Journal of the Royal Statistical Society: Series B (Methodological) , volume=

    Optimum experimental designs , author=. Journal of the Royal Statistical Society: Series B (Methodological) , volume=. 1959 , publisher=

    1959

  78. [78]

    The American Statistician , volume=

    Correlation in polynomial regression , author=. The American Statistician , volume=. 1979 , publisher=

    1979

  79. [79]

    Budescu , title =

    David V. Budescu , title =. Multivariate Behavioral Research , volume =. 1980 , publisher =. doi:10.1207/s15327906mbr1504\_7 , note =

    work page doi:10.1207/s15327906mbr1504 1980

  80. [80]

    Ruberg and James W

    Stephen J. Ruberg and James W. Stegeman , journal =. Pooling Data for Stability Studies: Testing the Equality of Batch Degradation Slopes , urldate =

Showing first 80 references.