# How Transparent and Reproducible Are Studies That Use Animal Models of Opioid Addiction?

## TL;DR

This study evaluates how transparent and reproducible research on animal models of opioid addiction has been from 2019 to 2023.

## Contribution

The study provides a comprehensive assessment of transparency and reproducibility practices in preclinical opioid addiction research.

## Key findings

- Transparency measures like preregistration and open data sharing were rarely used in the studied articles.
- Most papers failed to implement proper bias minimization practices and sample size calculations.
- About half of the papers had p-value inconsistencies, and 11% had statistical significance errors.

## Abstract

The reproducibility crisis in psychology has caused various fields to consider the reliability of their own findings. Many of the unfortunate aspects of research design that undermine reproducibility also threaten translation potential. In preclinical addiction research, the rates of translation have been disappointing. We tallied indices of transparency and accurate and thorough reporting in animal models of opioid addiction from 2019 to 2023. By examining the prevalence of these practices, we aimed to understand whether efforts to improve reproducibility are relevant to this field. For 255 articles, we report the prevalence of transparency measures such as preregistration, registered reports, open data and open code, as well as compliance to the Animal Research: Reporting of In Vivo Experiments (ARRIVE) guidelines. We also report rates of bias minimization practices (randomization, masking and data exclusion), sample size calculations and multiple corrections adjustments. Lastly, we estimated the accuracy of test statistic reporting using a version of StatCheck. All the transparency measures and the ARRIVE guideline items had low prevalence, including no cases of study preregistration and no cases where authors shared their analysis code. Similarly, the levels of bias minimization practices and sample size calculations were unsatisfactory. In contrast, adjustments for multiple comparisons were implemented in most articles (76.5%). Lastly, p‐value inconsistencies with test statistics were detected in about half of papers, and 11% contained statistical significance errors. We recommend that researchers, journal editors and others take steps to improve study reporting and to facilitate both replication and translation.

This study assessed transparency, reproducibility and bias minimization practices in animal models of opioid addiction research from 2019 to 2023. Key indices included analysis script sharing, preregistration, masking and randomization, among others. Our findings reveal substantial room for improvement to enhance transparency, reproducibility and the translational potential of pre‐clinical addiction research.