Detecting and Adjusting for Publication Bias in Meta-Analysis: A landscape review
Year: 2026
Type: article
Abstract: Companion code and data for the manuscript "Detecting and Adjusting for Publication Bias in Meta-Analysis: A landscape review" (Cattaneo, Santos, Albanese, Di Tanna), submitted to Cochrane Evidence Synthesis and Methods. The repository contains an R Markdown tutorial (Landscape_Review.Rmd) and its knitted HTML output that illustrate, with two worked examples, how to detect and adjust for publication bias in a random-effects meta-analysis: funnel plot-based approaches, small-study effects adjustments (including PET-PEESE and limit meta-analysis), methods based on the distribution of significant p-values, explicit selection models, power-based methods, and Bayesian model averaging (RoBMA and RoBMA-PSMA). The examples use study-level summary statistics extracted from two published meta-analyses: Deng and Wang (2024), k = 18, and García-Correa et al. (2021), pain intensity subset, k = 15 (data/deng_meta.xlsx and data/garcia_meta.xlsx). Please cite the original meta-analyses when reusing the data. To reproduce the analysis, install R (4.5.3 was used), JAGS and the packages listed in the README, open the project in RStudio so that here() resolves paths correctly, and knit Landscape_Review.Rmd. The RoBMA-PSMA fits are the slowest step and can take several minutes; they are fixed with seed = 1. Licences: code under MIT; text and knitted output under CC BY 4.0; data under the terms of the original publications.
Language: English
Subfield: Statistics, Probability and Uncertainty
Field: Decision Sciences
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7225425791