Abstract

Meta-analysis upweights studies reporting lower standard errors and hence more precision. But in observational settings common to much research on human behavior, precision is not given to the researcher. Precision must be estimated, and thus can be p-hacked to achieve statistical significance. Simulations and large-scale empirical applications show that spurious precision can invalidate inverse-variance weighting and bias-correction methods based on the funnel plot. Selection models fail to solve the problem, and common cures to publication bias can become worse than the disease. We introduce an approach (Meta-Analysis Instrumental Variable Estimator, MAIVE) that addresses spurious precision and limits the resulting bias in meta-analysis.

One-click meta-analysis in your browser: EasyMeta.org How to run MAIVE, on real data
Spurious precision
Fig: Spurious precision (right panel) renders common meta-analysis techniques biased

Reference: Zuzana Irsova, Pedro Bom, Tomas Havranek, and Heiko Rachinger (2025): "Spurious Precision in Meta-Analysis of Observational Research." Nature Communications 16, 8454. doi.org/10.1038/s41467-025-63261-0

Read the paper

Headline result

Spurious precision in meta-analysis (the MAIVE estimator): the methodological finding is estimated precision can be p-hacked, which can undermine inverse-variance weighting; MAIVE instruments variance with sample size (Irsova et al. 2025, Nature Communications).

How to cite

Zuzana Irsova, Pedro Bom, Tomas Havranek, and Heiko Rachinger (2025): “Spurious Precision in Meta-Analysis of Observational Research.” Nature Communications 16, 8454. doi.org/10.1038/s41467-025-63261-0

BibTeX 
@article{irsova2025maive,
  author  = {Zuzana Irsova and Pedro Bom and Tomas Havranek and Heiko Rachinger},
  title   = {Spurious Precision in Meta-Analysis of Observational Research},
  journal = {Nature Communications},
  year    = {2025},
  doi     = {10.1038/s41467-025-63261-0},
}

Extensions

MAIVE is the published estimator and what this page is about. Two methods build on it. Neither is finished or has a paper yet; the talks below are the only source.

WAIVE, the Weighted Adjustment Instrumental Variable Estimator, is the more aggressive correction. MAIVE instruments the standard error with sample size; WAIVE downweights the excess precision that is left.

The residual discontinuity test is a test for p-hacking built on the same first stage. Excess precision, the part of a standard error that sample size does not explain, should not jump at the significance threshold. RDT asks whether it does.

Both are experimental options in EasyMeta, beside MAIVE. The talks are MAER-Net, Ottawa, 2025 on WAIVE, SRSM, Chania, 2026, and MAER-Net, Chemnitz, 2026 on the test.