# Mitigating LLM-based p-Hacking by Preregistering for the Next LLM

arXiv: https://arxiv.org/abs/2606.27687 · HTML: https://mlsanity.com/paper/2606.27687/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 26 of its 53 references that are in the map.
- First posted on arXiv 26 Jun 2026.

## Builds on (most cited first, 26 of 26)

- [Out of One, Many: Using Language Models to Simulate Human Samples](https://mlsanity.com/paper/2209.06899/) (2022)
- [Agent Laboratory: Using LLM Agents as Research Assistants](https://mlsanity.com/paper/2501.04227/) (2025)
- [The AI Scientist-v2: Workshop-Level Automated Scientific Discovery via Agentic Tree Search](https://mlsanity.com/paper/2504.08066/) (2025)
- [Can Large Language Models Transform Computational Social Science?](https://mlsanity.com/paper/2305.03514/) (2023)
- [Prediction-Powered Inference](https://mlsanity.com/paper/2301.09633/) (2023)
- [Causal Inference in Natural Language Processing: Estimation, Prediction, Interpretation and Beyond](https://mlsanity.com/paper/2109.00725/) (2021)
- [The Effect of Sampling Temperature on Problem Solving in Large Language Models](https://mlsanity.com/paper/2402.05201/) (2024)
- [Towards End-to-End Automation of AI Research](https://mlsanity.com/paper/2606.15497/) (2026)
- [Language Model Fine-Tuning on Scaled Survey Data for Predicting Distributions of Public Opinions](https://mlsanity.com/paper/2502.16761/) (2025)
- [Large Language Model Hacking: Quantifying the Hidden Risks of Using LLMs for Text Annotation](https://mlsanity.com/paper/2509.08825/) (2025)
- [Can Unconfident LLM Annotations Be Used for Confident Conclusions?](https://mlsanity.com/paper/2408.15204/) (2024)
- [Using Imperfect Surrogates for Downstream Inference: Design-based Supervised Learning for Social Science Applications of Large Language Models](https://mlsanity.com/paper/2306.04746/) (2023)
- [Finetuning LLMs for Human Behavior Prediction in Social Science Experiments](https://mlsanity.com/paper/2509.05830/) (2025)
- [Flaw or Artifact? Rethinking Prompt Sensitivity in Evaluating LLMs](https://mlsanity.com/paper/2509.01790/) (2025)
- [The More You Automate, the Less You See: Hidden Pitfalls of AI Scientist Systems](https://mlsanity.com/paper/2509.08713/) (2025)
- [This human study did not involve human subjects: Validating LLM simulations as behavioral evidence](https://mlsanity.com/paper/2602.15785/) (2026)
- [Towards LLMs Robustness to Changes in Prompt Format Styles](https://mlsanity.com/paper/2504.06969/) (2025)
- [Detecting LLM-Generated Peer Reviews](https://mlsanity.com/paper/2503.15772/) (2025)
- [Valid Survey Simulations with Limited Human Data: The Roles of Prompting, Fine-Tuning, and Rectification](https://mlsanity.com/paper/2510.11408/) (2025)
- [An Empirical Study of LLM Reasoning Ability Under Strict Output Length Constraint](https://mlsanity.com/paper/2504.14350/) (2025)
- [Prompt Design Matters for Computational Social Science Tasks but in Unpredictable Ways](https://mlsanity.com/paper/2406.11980/) (2024)
- [Usefulness of LLMs as an Author Checklist Assistant for Scientific Papers: NeurIPS'24 Experiment](https://mlsanity.com/paper/2411.03417/) (2024)
- [Doubly-Robust LLM-as-a-Judge: Externally Valid Estimation with Imperfect Personas](https://mlsanity.com/paper/2509.22957/) (2025)
- [Stop Drawing Scientific Claims from LLM Social Simulations Without Robustness Audits](https://mlsanity.com/paper/2605.18890/) (2026)
- [A Randomized Controlled Trial on Anonymizing Reviewers to Each Other in Peer Review Discussions](https://mlsanity.com/paper/2403.01015/) (2024)
- [Policies Permitting LLM Use for Polishing Peer Reviews Are Currently Not Enforceable](https://mlsanity.com/paper/2603.20450/) (2026)

## Built on by (most cited first, 2 of 2)

- [Certifying Compressed Language Models: An Audit and a Statistical Toolkit](https://mlsanity.com/paper/2608.15046/) (2026)
- [What Does a Harness Repair? A Preregistered Study of Visibility, Baseline Adequacy and Evaluation Defects](https://mlsanity.com/paper/2610.05533/) (2026)

## Abstract (arXiv)

Large language models (LLMs) are increasingly used to generate, classify, and annotate data whose outputs feed downstream hypothesis tests. However, LLM-based research is easy to p-hack: a researcher can tune the prompts, decoding parameters, or output format until a desired result is reached. We propose a protocol to mitigate p-hacking in LLM-based research: preregistering the experiment and eligible models, and then running it on the first eligible LLM that is released after the preregistration. The researcher finalizes the procedure on current models, preregisters the analysis plan together with a set of eligible future models, and runs the confirmatory analysis on the first eligible model released afterward. Because this model does not exist at commitment time, it cannot be hacked against; furthermore, configurations that hack one model frequently do not transfer to the next. We evaluate the protocol on two tasks whose true values are known. Across 20 models from four providers and 11 LLM-analysis configurations, the protocol would have blocked successful transfer of the p-hack in 73.9% and 72.7% of cases in the two tasks. Additional analyses reveal that mitigation remains substantial under several stress tests. Finally, putting money where our mouth is, we followed our own protocol and preregistered our experiment. The preregistered experiment confirmed the protocol's effectiveness: out of the 7 configurations that hacked the prior model, the hacking failed to carry over in 6 configurations on the first eligible model released afterward.