EIOSE - ANR

Principal investigator

Thibaut Arpinon - Chair Professor Junior (CPJ) at Savoie Mont-Blanc University (USMB) and IREGE.  

Project team 

Ph.D. student to be recruited soon...
Post-doctoral researcher to be recruited soon...

Scientific contributors and co-authors

To be added soon...

In the last decade, social sciences have seen the emergence of a scientific movement around a fundamental question: how much can we trust scientific results? Recent developments, such as the replication crisis suspicion (Loken & Gelman, 2017), uncovering of fraudulent data (Callaway, 2011; Simmons, 2024), and identification of questionable research practices (Wagenmakers et al., 2011) have shed light on problematic scientific practices. As a result, social scientists have developed an array of tools, under the umbrella term Open Science (Pennington, 2023), to unify practices, increase transparency, and ensure the credibility and replicability of findings. However, despite a call for more transparency, the transition remains slow (Nosek et al., 2015), and more efforts are needed.

The field of economics is not exempt. Seminal work suggests that not all economic results replicate successfully (Camerer et al., 2016) and that economic research does not self-correct (Ankel-Peters et al., 2024). Alongside, questionable research and publishing practices have been uncovered. On the one hand, publication bias (i.e., favoring positive results) is widespread (Chopra et al., 2024) and is supported by the publish-or-perish culture. On the other hand, several papers have identified questionable research practices such as p-hacking (Brodeur et al., 2016) or underpowered testing (Ioannidis et al., 2017), that lead to exaggerated effect sizes, false positive results, and alter the credibility of findings. As a result, economists have been calling for more replications (Page et al., 2021) and have launched large-scale replication initiatives (e.g., I4R - Institute For Replications). 

Unfortunately, although replications and identification of malpractices are useful ex-post tools to raise awareness, their ex-ante impact remains unknown. In response, pre-registration and Registered Reports have been introduced. Pre-registration is a valuable tool to distinguish predictions from postdictions (Nosek et al., 2018) and its popularity is rapidly increasing in economics (Arpinon, 2025; Arpinon & Lefebvre, 2024). Nevertheless, pre-registration is relatively new in economics, and further developments are required to validate its effectiveness (e.g., scientific evaluation and evidence). In parallel, Registered Reports were launched to counter publication bias (Chambers & Tzavella, 2022). The format’s unique reviewing process secures the publication before the results are known and eliminates incentives to find positive results. Even though Registered Reports are powerful, their adoption in economics is also relatively recent (Arpinon & Espinosa, 2023) and efforts are still required to eliminate perceived barriers.

The hypothesis driving this ANR project is that the transition towards better open science practices in the field of economics is hindered by insufficient evidence on malpractices, misperceptions, and current unadapted tools. The growing records of questionable practices call for further research, better understanding, and methodological advancements. In particular, the greatest challenge is to fit current improvements to the needs of economists without imposing drastic changes or altering research quality. In turn, equipping economic researchers with better tools will relieve some of the burden off their shoulders and will allow for a smoother and quicker implementation of essential measures. Therefore, the objective is to accelerate the development of open science practices in economics by contributing to the accumulation of evidence on malpractices, better understanding economists’ current perceptions and needs, and facilitating open science implementation by improving on current tools and developing newer and more adapted ones.

The EIOSE project is divided into 3 distinct work-packages.

In this work package, we propose to empirically assess the extent of questionable publishing and research practices in economics and theoretically characterize the mechanisms at play.  We will put a strong emphasis on identifying p-hacking under unexplored facets to unveil hidden drivers and will characterize how the misuse of current open science practices leads to questionable practices. The goal is to offer a comprehensive view of questionable practices in the economics literature to effectively inform policymakers for future open science developments.

In this work package, we hypothesize that the low adoption rates and heterogeneity in the use of open science in economics stem from misbeliefs, misconceptions, and misperception. We propose to explore this facet by empirically assessing current practices, perceptions, and beliefs to better understand economists’ expectations and correct potential misbeliefs. The overarching goal is to obtain a broad picture of economists’ needs to tailor current and future practices and maximize implementation to foster credible results.

This work package hypothesizes that the misuse, slow adoption rate, and reluctance to fully embrace open science practices stem from inefficient and unadapted tools. We propose to improve on current and develop new tools to facilitate implementation and increase adoption rates. The broader goal is to facilitate open science tools’ use to maximize their benefits, and thus improve research credibility.