Pith. sign in

REVIEW 1 minor 6 references

regcorr: An R Package for Regression Models of Pearson Correlation Coefficients

T0 review · 0 major / 1 minor · reviewed 2026-06-28 · grok-4.3

Pith's one-line read The regcorr package provides regression models that link Pearson correlation coefficients to covariates for bivariate responses.

desk verdict This is a software paper that packages the 2023 Dufera et al. likelihood framework into a lightweight R package for regressing bivariate correlations on covariates. read the letter →

arxiv 2606.05676 v1 pith:VO22R4EV submitted 2026-06-04 stat.ME stat.CO

classification stat.MEstat.CO
keywords PearsoncorrelationregressionmodelingRpackagecovariateeffectsbivariateresponseslikelihoodestimationbootstrapinference
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper presents an R package that treats the Pearson correlation between two responses as a quantity that can itself be regressed on covariates. Users can therefore examine how factors such as demographics or experimental conditions change the strength of association rather than assuming the correlation is fixed. The implementation covers both bivariate normal data and bivariate binary data, supplies Newton-Raphson estimation, includes functions to generate simulation data, and adds a bootstrap routine to assess whether covariate effects on correlation are significant. Readers would care because many scientific questions concern not only average levels but also when and why relationships between variables become stronger or weaker.

What carries the argument

The regression link that expresses the correlation parameter as a function of covariates, maximized via Newton-Raphson on the bivariate likelihood.

What would settle it

Simulate data from the model with known nonzero covariate effects on correlation, then check whether the package recovers those effects and their significance at the rates implied by the bootstrap procedure.

Watch

Extended reading notes

Core claim

The paper claims that the regcorr package implements a likelihood-based regression in which the Pearson correlation coefficient is expressed as a function of a linear predictor of covariates, supports bivariate normal and bivariate Bernoulli responses, performs estimation via Newton-Raphson, supplies data generators, and provides bootstrap-based inference for the significance and power of covariate effects on the correlation.

Load-bearing premise

The likelihood framework accurately describes how covariates affect the correlation parameter and the package implements that framework without errors.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

0 major / 1 minor

Summary. The manuscript describes the regcorr R package, which implements regression models linking a Pearson correlation coefficient to a linear predictor of covariates. It supports bivariate normal and bivariate Bernoulli responses, uses Newton-Raphson for maximum-likelihood estimation, includes data generators for simulations, and provides a bootstrap subroutine for assessing covariate effects on correlations. The implementation follows the likelihood framework of Dufera, Liu, and Xu (2023) and is distributed on CRAN with minimal dependencies and no compiled code.

Significance. If the implementation faithfully realizes the cited framework, the package supplies a lightweight, accessible tool for modeling heterogeneous associations in bivariate data. The inclusion of simulation generators and bootstrap routines directly supports reproducibility of power and significance assessments, which is a practical strength for applied work in fields where correlation strength varies with covariates.

minor comments (1)
  1. The abstract states that the package 'exposes' the Dufera-Liu-Xu framework; a brief explicit statement in the computational-design section confirming that the score equations and Hessian match those in the 2023 reference would strengthen the claim of faithful implementation.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for their positive assessment of the manuscript, the accurate summary of the regcorr package, and the recommendation to accept. We are pleased that the practical strengths for applied work in modeling heterogeneous correlations were recognized.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; external framework implemented

full rationale

The paper is a software description whose core contribution is an R package exposing the likelihood framework of Dufera, Liu, and Xu (2023). This is an external citation with non-overlapping authors; the package adds Newton-Raphson routines, data generators, and bootstrap tools but does not derive or redefine the underlying statistical model. No equations, ansatzes, or uniqueness claims reduce to self-citation chains or fitted inputs by construction. The derivation chain is therefore self-contained against the cited external reference.

Assumptions & free parameters 0 free parameters · 0 assumptions · 0 invented entities

No new free parameters, axioms, or invented entities introduced; relies on standard likelihood methods from prior work.

how reviews work

0 comments
Cite this review

Pith. "Pith review of regcorr: An R Package for Regression Models of Pearson Correlation Coefficients." pith.science (2026). https://pith.science/paper/VO22R4EV

@misc{pith2026260605676,
  author       = {Pith},
  title        = {Pith review of: regcorr: An R Package for Regression Models of Pearson Correlation Coefficients},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/VO22R4EV}},
  note         = {Machine review of arXiv:2606.05676}
}
read the original abstract

Pearson's correlation coefficient is commonly used as a single-number summary of association between two responses. In many applications, however, the strength of association is itself heterogeneous and may vary with demographic, biological, experimental, or environmental covariates. The regcorr package implements regression models in which a Pearson correlation coefficient is linked to a linear predictor of covariates. The package supports bivariate normal responses and bivariate Bernoulli responses, provides Newton-Raphson estimation routines, includes data generators for simulation studies, and supplies a bootstrap-based subroutine for assessing the significance and power of covariate effects. The implementation follows the likelihood-based framework of Dufera, Liu, and Xu (2023) and exposes it through a lightweight R interface with no compiled code and minimal dependencies. This paper describes the statistical model, the computational design of regcorr, reproducible usage examples, and practical guidance for interpreting covariate-dependent correlations. The package is available from the Comprehensive R Archive Network at https://CRAN.R-project.org/package=regcorr under the MIT license.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

6 extracted references · 1 canonical work pages

  1. [1]

    G., Liu, T., and Xu, J

    Dufera, A. G., Liu, T., and Xu, J. (2023). Regression mode ls of Pearson correlation coefficient. Statistical Theory and Related Fields , 7(2), 97–106. doi:10.1080/24754269.2023.2164970

  2. [2]

    Bartlett, R. F. (1993). Linear modelling of Pearson’s pr oduct moment correlation coefficient: An application of Fisher’s z-transformation. Journal of the Royal Statistical Society: Series D, 42(1), 45–53

  3. [3]

    Liu, Q., Li, C., Wang, V., and Shepherd, B. E. (2018). Cova riate-adjusted Spearman rank correlation with probability-scale residuals. Biometrics, 74(2), 595–605

  4. [4]

    Qaqish, B. F. (2003). A family of multivariate binary dis tributions for simulating correlated binary variables with specified marginal means and correlat ions. Biometrika, 90(2), 455–463

  5. [5]

    K., and McFadden, D

    Newey, W. K., and McFadden, D. (1994). Large sample estim ation and hypothesis testing. In Handbook of Econometrics , volume 4, 2111–2245. Elsevier

  6. [6]

    R: A Language and Environment for Statistical Computing

    R Core Team. R: A Language and Environment for Statistical Computing . R Foundation for Statistical Computing, Vienna, Austria. https://www.R-project.org/. 8

Pith tools

Reviewed June 28, 2026 · model on record in the stance chip above.