Band 9, Heft 2

Purpose Assess whether ChatGPT 4.0 is accurate enough to perform research evaluations on journal articles to automate this time-consuming task. Design/methodology/approach Test the extent to which ChatGPT-4 can assess the quality of journal articles using a case study of the published scoring guidelines of the UK Research Excellence Framework (REF) 2021 to create a research evaluation ChatGPT. This was applied to 51 of my own articles and compared against my own quality judgements. Findings ChatGPT-4 can produce plausible document summaries and quality evaluation rationales that match the REF criteria. Its overall scores have weak correlations with my self-evaluation scores of the same documents (averaging r=0.281 over 15 iterations, with 8 being statistically significantly different from 0). In contrast, the average scores from the 15 iterations produced a statistically significant positive correlation of 0.509. Thus, averaging scores from multiple ChatGPT-4 rounds seems more effective than individual scores. The positive correlation may be due to ChatGPT being able to extract the author’s significance, rigour, and originality claims from inside each paper. If my weakest articles are removed, then the correlation with average scores (r=0.200) falls below statistical significance, suggesting that ChatGPT struggles to make fine-grained evaluations. Research limitations The data is self-evaluations of a convenience sample of articles from one academic in one field. Practical implications Overall, ChatGPT does not yet seem to be accurate enough to be trusted for any formal or informal research quality evaluation tasks. Research evaluators, including journal editors, should therefore take steps to control its use. Originality/value This is the first published attempt at post-publication expert review accuracy testing for ChatGPT.

Purpose Recently, global science has shown an increasing open trend, however, the characteristics of research integrity of open access (OA) publications have rarely been studied. The aim of this study is to compare the characteristics of retracted articles across different OA levels and discover whether OA level influences the characteristics of retracted articles. Design/methodology/approach The research conducted an analysis of 6,005 retracted publications between 2001 and 2020 from the Web of Science and Retraction Watch databases. These publications were categorized based on their OA levels, including Gold OA, Green OA, and non-OA. The study explored retraction rates, time lags and reasons within these categories. Findings The findings of this research revealed distinct patterns in retraction rates among different OA levels. Publications with Gold OA demonstrated the highest retraction rate, followed by Green OA and non-OA. A comparison of retraction reasons between Gold OA and non-OA categories indicated similar proportions, while Green OA exhibited a higher proportion due to falsification and manipulation issues, along with a lower occurrence of plagiarism and authorship issues. The retraction time lag was shortest for Gold OA, followed by non-OA, and longest for Green OA. The prolonged retraction time for Green OA could be attributed to an atypical distribution of retraction reasons. Research limitations There is no exploration of a wider range of OA levels, such as Hybrid OA and Bronze OA. Practical implications The outcomes of this study suggest the need for increased attention to research integrity within the OA publications. The occurrences of falsification, manipulation, and ethical concerns within Green OA publications warrant attention from the scientific community. Originality/value This study contributes to the understanding of research integrity in the realm of OA publications, shedding light on retraction patterns and reasons across different OA levels.

Purpose The notable increase in retraction papers has attracted considerable attention from diverse stakeholders. Various sources are now offering information related to research integrity, including concerns voiced on social media, disclosed lists of paper mills, and retraction notices accessible through journal websites. However, despite the availability of such resources, there remains a lack of a unified platform to consolidate this information, thereby hindering efficient searching and cross-referencing. Thus, it is imperative to develop a comprehensive platform for retracted papers and related concerns. This article aims to introduce “Amend,” a platform designed to integrate information on research integrity from diverse sources. Design/methodology/approach The Amend platform consolidates concerns and lists of problematic articles sourced from social media platforms (e.g., PubPeer, For Better Science), retraction notices from journal websites, and citation databases (e.g., Web of Science, CrossRef). Moreover, Amend includes investigation and punishment announcements released by administrative agencies (e.g., NSFC, MOE, MOST, CAS). Each related paper is marked and can be traced back to its information source via a provided link. Furthermore, the Amend database incorporates various attributes of retracted articles, including citation topics, funding details, open access status, and more. The reasons for retraction are identified and classified as either academic misconduct or honest errors, with detailed subcategories provided for further clarity. Findings Within the Amend platform, a total of 32,515 retracted papers indexed in SCI, SSCI, and ESCI between 1980 and 2023 were identified. Of these, 26,620 (81.87%) were associated with academic misconduct. The retraction rate stands at 6.64 per 10,000 articles. Notably, the retraction rate for non-gold open access articles significantly differs from that for gold open access articles, with this disparity progressively widening over the years. Furthermore, the reasons for retractions have shifted from traditional individual behaviors like falsification, fabrication, plagiarism, and duplication to more organized large-scale fraudulent practices, including Paper Mills, Fake Peer-review, and Artificial Intelligence Generated Content (AIGC). Research limitations The Amend platform may not fully capture all retracted and concerning papers, thereby impacting its comprehensiveness. Additionally, inaccuracies in retraction notices may lead to errors in tagged reasons. Practical implications Amend provides an integrated platform for stakeholders to enhance monitoring, analysis, and research on academic misconduct issues. Ultimately, the Amend database can contribute to upholding scientific integrity. Originality/value This study introduces a globally integrated platform for retracted and concerning papers, along with a preliminary analysis of the evolutionary trends in retracted papers.

Purpose The goal of this study is a comparative analysis of the relation between funding (a main driver for scientific research) and citations in papers of Nobel Laureates in physics, chemistry and medicine over 2019-2020 and the same relation in these research fields as a whole. Design/Methodology/Approach This study utilizes a power law model to explore the relationship between research funding and citations of related papers. The study here analyzes 3,539 recorded documents by Nobel Laureates in physics, chemistry and medicine and a broader dataset of 183,016 documents related to the fields of physics, medicine, and chemistry recorded in the Web of Science database. Findings Results reveal that in chemistry and medicine, funded researches published in papers of Nobel Laureates have higher citations than unfunded studies published in articles; vice versa high citations of Nobel Laureates in physics are for unfunded studies published in papers. Instead, when overall data of publications and citations in physics, chemistry and medicine are analyzed, all papers based on funded researches show higher citations than unfunded ones. Originality/Value Results clarify the driving role of research funding for science diffusion that are systematized in general properties: a) articles concerning funded researches receive more citations than (un)funded studies published in papers of physics, chemistry and medicine sciences, generating a high Matthew effect (a higher growth of citations with the increase in the number of papers); b) research funding increases the citations of articles in fields oriented to applied research (e.g., chemistry and medicine) more than fields oriented towards basic research (e.g., physics). Practical Implications The results here explain some characteristics of scientific development and diffusion, highlighting the critical role of research funding in fostering citations and the expansion of scientific knowledge. This finding can support decisionmaking of policymakers and R&D managers to improve the effectiveness in allocating financial resources in science policies to generate a higher positive scientific and societal impact.

Purpose Many science, technology and innovation (STI) resources are attached with several different labels. To assign automatically the resulting labels to an interested instance, many approaches with good performance on the benchmark datasets have been proposed for multilabel classification task in the literature. Furthermore, several open-source tools implementing these approaches have also been developed. However, the characteristics of real-world multilabel patent and publication datasets are not completely in line with those of benchmark ones. Therefore, the main purpose of this paper is to evaluate comprehensively seven multi-label classification methods on real-world datasets. Design/methodology/approach Three real-world datasets (Biological-Sciences, Health-Sciences, and USPTO) from SciGraph and USPTO database are constructed. Seven multilabel classification methods with tuned parameters (dependency-LDA, ML k NN, LabelPowerset, RA k EL, TextCNN, TexRNN, and TextRCNN) are comprehensively compared on these three real-world datasets. To evaluate the performance, the study adopts three classification-based metrics: Macro-F1, Micro-F1, and Hamming Loss. Findings The TextCNN and TextRCNN models show obvious superiority on small-scale datasets with more complex hierarchical structure of labels and more balanced documentlabel distribution in terms of macro-F1, micro-F1 and Hamming Loss. The ML k NN method works better on the larger-scale dataset with more unbalanced document-label distribution. Research limitations Three real-world datasets differ in the following aspects: statement, data quality, and purposes. Additionally, open-source tools designed for multi-label classification also have intrinsic differences in their approaches for data processing and feature selection, which in turn impacts the performance of a multi-label classification approach. In the near future, we will enhance experimental precision and reinforce the validity of conclusions by employing more rigorous control over variables through introducing expanded parameter settings. Practical implications The observed Macro F1 and Micro F1 scores on real-world datasets typically fall short of those achieved on benchmark datasets, underscoring the complexity of real-world multi-label classification tasks. Approaches leveraging deep learning techniques offer promising solutions by accommodating the hierarchical relationships and interdependencies among labels. With ongoing enhancements in deep learning algorithms and large-scale models, it is expected that the efficacy of multi-label classification tasks will be significantly improved, reaching a level of practical utility in the foreseeable future. Originality/value (1) Seven multi-label classification methods are comprehensively compared on three real-world datasets. (2) The TextCNN and TextRCNN models perform better on small-scale datasets with more complex hierarchical structure of labels and more balanced document-label distribution. (3) The ML k NN method works better on the larger-scale dataset with more unbalanced document-label distribution.

Research data infrastructures form the cornerstone in both cyber and physical spaces, driving the progression of the data-intensive scientific research paradigm. This opinion paper presents an overview of global research data infrastructure, drawing insights from national roadmaps and strategic documents related to research data infrastructure. It emphasizes the pivotal role of research data infrastructures by delineating four new missions aimed at positioning them at the core of the current scientific research and communication ecosystem. The four new missions of research data infrastructures are: (1) as a pioneer, to transcend the disciplinary border and address complex, cutting-edge scientific and social challenges with problem- and data-oriented insights; (2) as an architect, to establish a digital, intelligent, flexible research and knowledge services environment; (3) as a platform, to foster the high-end academic communication; (4) as a coordinator, to balance scientific openness with ethics needs.