Accountability in Research-Policies and Quality Assurance最新文献

Research misconduct, broadly defined as acts of fabrication, falsification and/or plagiarism, violate the value system of science, cost significant wastage of public resources, and in more extreme cases endanger research participants or members of the society at large. Determination of culpability in research misconduct requires establishment of intent on the part of the respondent or perpetrator. However, "intent" is a state of mind, and its perception is subjective, unequivocal evidence for which would not be as readily established compared to the objective evidence available for the acts themselves. Here, we explore the concept of "intent" in research misconduct, how it is framed in criminological/legal terms, and narrated from a psychological perspective. Based on these, we propose a framework whereby lines of questioning and investigation, as defined by legislative terms and informed by the models and tools of psychology, could help in establishing a preponderance of evidence for culpable intent. Such a framework could be useful in research misconduct adjudications and in delivering sanctions.

Artificial Intelligence (AI) language models continue to expand in both access and capability. As these models have evolved, the number of academic journals in medicine and healthcare which have explored policies regarding AI-generated text has increased. The implementation of such policies requires accurate AI detection tools. Inaccurate detectors risk unnecessary penalties for human authors and/or may compromise the effective enforcement of guidelines against AI-generated content. Yet, the accuracy of AI text detection tools in identifying human-written versus AI-generated content has been found to vary across published studies. This experimental study used a sample of behavioral health publications and found problematic false positive and false negative rates from both free and paid AI detection tools. The study assessed 100 research articles from 2016-2018 in behavioral health and psychiatry journals and 200 texts produced by AI chatbots (100 by "ChatGPT" and 100 by "Claude"). The free AI detector showed a median of 27.2% for the proportion of academic text identified as AI-generated, while commercial software Originality.AI demonstrated better performance but still had limitations, especially in detecting texts generated by Claude. These error rates raise doubts about relying on AI detectors to enforce strict policies around AI text generation in behavioral health publications.

Resnik, Hosseini and Rasmussen's take on universities having conflict of interest (COI), and should engage an independent research integrity organization in investigating research misconduct (RM) allegations against top officials, is prudent and timely. COI could be conceived either against or for in the processes toward the conviction of a respondent. For the latter, we need look no further than another recent Harvard case involving Francesca Gino.

Recent developments in the German academic landscape have seen a shifting approach to promoting research integrity. In 2019, the German Research Foundation (DFG) incentivized all research and higher education institutions to appoint ombudspersons who advise members of their institution in matters of good research practice or suspected research misconduct. These ombudspersons for good research practice, usually professors who act in this function on a voluntary basis, need institutional support to be prepared for and fulfill their diverse duties. The Ombuds-Modelle@BUA (2020) and OBUA - Ombudswesen@BUA (2021-2023) projects worked to advance the professionalization of ombudspersons in the Berlin research area by first investigating the current situation and then offering a meta-level of support in training, networking, and knowledge exchange. Furthermore, the OBUA project engaged in meta-research, investigating the status quo of local ombuds systems and demands for support. The project findings, discussed in this contribution, show that the professionalization of local ombuds systems has been evolving in past years, especially in the areas of training and networking. Infrastructural support measures, however, remain largely underdeveloped.

The objective of this scoping review was to systematically review the literature on how non-financial conflicts of interest (nfCOI) are defined and evaluated, and the strategies suggested for their management in health-related and biomedical journals. PubMed, Embase, Scopus and Web of Science were searched for peer reviewed studies published in English between 1970 and December 2023 that addressed at least one of the following: the definition, evaluation, or management of non-financial conflicts of interest. From 658 studies, 190 studies were included in the review. nfCOI were discussed most commonly in empirical (22%; 42/190), theoretical (15%; 29/190) and "other" studies (18%; 34/190) - including commentary, perspective, and opinion articles. nfCOI were addressed frequently in the research domain (36%; 68/190), publication domain (29%; 55/190) and clinical practice domain (17%; 32/190). Attitudes toward nfCOI and their management were divided into two distinct groups. The first larger group claimed that nfCOI were problematic and required some form of management, whereas the second group argued that nfCOI were not problematic, and therefore, did not require management. Despite ongoing debates about the nature, definition, and management of nfCOI, many articles included in this review agreed that serious consideration needs to be given to the prevalence, impact and optimal mitigation of non-financial COI.

This paper investigates the differences and similarities between European regulatory research integrity systems. The data collection process involved gathering information from public sources. A total of 27 European countries were included in the comprehensive dataset. Three determinants were examined: the legal structure of national research integrity regulation, the presence of national research integrity guidelines, and the provision of research integrity training by national research integrity offices. Qualitative content analysis was employed to identify relevant differences in national research integrity systems and the work of national research integrity offices. The findings suggest that the functions and powers of research integrity offices in Europe vary significantly, and there is extensive variation in the legal status and functions of national research integrity systems. We identify the major implications arising from these differences and explore what the challenges for harmonization of the European research integrity systems are. Our findings highlight the need for promoting dialogue between actors on an international level.

Previous studies have found that factors such as gender and academic positions do not influence the severity of administrative actions taken by institutions. However, this study provides partly inconsistent evidence. It focuses on incidents of research misconduct in hospitals across Mainland China and explores factors related to punishment using a large cross-sectional dataset (N = 815). Regression analysis revealed a significant correlation between authorship order and the punishment intensity (p < 0.05). Under specific conditions, there was a significant correlation between the professional title (senior) and punishment intensity (p = 0.001), and an interaction between professional title and types of research misbehavior. Further analysis of simple effects showed that, in cases of fabrication and falsification, and combinations of multiple research misbehavior, researchers with senior titles received significantly lighter punishments compared to those with junior, intermediate, and associate senior titles (p < 0.05). The study unveils the potential accountability patterns (collective accountability and tiered punishment) that may be adopted by hospitals in Mainland China, as well as the challenges faced in ensuring fairness, emphasizing the importance of independent investigative bodies for incidents of research misconduct, and advocating for fairness as a priority in governance of research misconduct.

Amid the COVID-19 pandemic, while the world sought solutions, few scholars exploited the situation for personal gains through deceptive studies and manipulated data. This paper presents the extent of 400 retracted COVID-19 papers listed by the RetractionWatch database until the month of February 2024. The primary purpose of the research was to analyze journal quality and retractions trends. Evaluating the journal's quality is vital for stakeholders, as it enables them to effectively address and prevent such incidents and their future repercussions. The present study found that one-fourth of publications were retracted within the first month of their publication, followed by an additional 6% within six months of publication. One third of the retractions originated from Q1 journals, with another significant portion coming from Q2 (29.8%). An analysis of the reasons for retractions indicates that a quarter of retractions were attributed to multiple causes, predominantly associated with publications in Q2 journals, while another quarter were linked to data issues, primarily observed in Q1 publications. Elsevier retracted 31% of papers, with the majority published as Q1, followed by Springer (11.5%), predominantly as Q2. The study also examined author contributions, revealing that 69.3% were male, with females (30.7%) mainly holding middle author positions.

Drawing on our experiences conducting replications we describe the lessons we learned about replication studies and formulate recommendations for researchers, policy makers, and funders about the role of replication in science and how it should be supported and funded. We first identify a variety of benefits of doing replication studies. Next, we argue that it is often necessary to improve aspects of the original study, even if that means deviating from the original protocol. Thirdly, we argue that replication studies highlight the importance of and need for more transparency of the research process, but also make clear how difficult that is. Fourthly, we underline that it is worth trying out replication in the humanities. We finish by formulating recommendations regarding reproduction and replication research, aimed specifically at funders, editors and publishers, and universities and other research institutes.