Purpose: Interdisciplinary research has become a critical approach to addressing complex societal, economic, technological, and environmental challenges, driving innovation and integrating scientific knowledge. While interdisciplinarity indicators are widely used to evaluate research performance, the impact of classification granularity on these assessments remains underexplored.
Design/methodology/approach: This study investigates how different levels of classification granularity—macro, meso, and micro—affect the evaluation of interdisciplinarity in research institutes. Using a dataset of 262 institutes from four major German non-university organizations (FHG, HGF, MPG, WGL) from 2018 to 2022, we examine inconsistencies in interdisciplinarity across levels, analyze ranking changes, and explore the influence of institutional fields and research focus (applied vs. basic).
Findings: Our findings reveal significant inconsistencies in interdisciplinarity across classification levels, with rankings varying substantially. Notably, the Fraunhofer Society (FHG), which performs well at the macro level, experiences significant ranking declines at meso and micro levels. Normalizing interdisciplinarity by research field confirmed that these declines persist. The research focus of institutes, whether applied, basic, or mixed, does not significantly explain the observed ranking dynamics.
Research limitations: This study has only considered the publication-based dimension of institutional interdisciplinarity and has not explored other aspects.
Practical implications: The findings provide insights for policymakers, research managers, and scholars to better interpret interdisciplinarity metrics and support interdisciplinary research effectively.
Originality/value: This study underscores the critical role of classification granularity in interdisciplinarity assessment and emphasizes the need for standardized approaches to ensure robust and fair evaluations.

Purpose: This research addresses the challenge of concept drift in AI-enabled software, particularly within autonomous vehicle systems where concept drift in object recognition (like pedestrian detection) can lead to misclassifications and safety risks. This study introduces a proactive framework to detect early signs of domain-specific concept drift by leveraging domain analysis and natural language processing techniques. This method is designed to help maintain the relevance of domain knowledge and prevent potential failures in AI systems due to evolving concept definitions.
Design/methodology/approach: The proposed framework integrates natural language processing and image analysis to continuously update and monitor key domain concepts against evolving external data sources, such as social media and news. By identifying terms and features closely associated with core concepts, the system anticipates and flags significant changes. This was tested in the automotive domain on the pedestrian concept, where the framework was evaluated for its capacity to detect shifts in the recognition of pedestrians, particularly during events like Halloween and specific car accidents.
Findings: The framework demonstrated an ability to detect shifts in the domain concept of pedestrians, as evidenced by contextual changes around major events. While it successfully identified pedestrian-related drift, the system’s accuracy varied when overlapping with larger social events. The results indicate the model’s potential to foresee relevant shifts before they impact autonomous systems, although further refinement is needed to handle high-impact concurrent events.
Research limitations: This study focused on detecting concept drift in the pedestrian domain within autonomous vehicles, with results varying across domains. To assess generalizability, we tested the framework for airplane-related incidents and demonstrated adaptability. However, unpredictable events and data biases from social media and news may obscure domain-specific drifts. Further evaluation across diverse applications is needed to enhance robustness in evolving AI environments.
Practical implications: The proactive detection of concept drift has significant implications for AI-driven domains, especially in safety-critical applications like autonomous driving. By identifying early signs of drift, this framework provides actionable insights for AI system updates, potentially reducing misclassification risks and enhancing public safety. Moreover, it enables timely interventions, reducing costly and labor-intensive retraining requirements by focusing only on the relevant aspects of evolving concepts. This method offers a streamlined approach for maintaining AI system performance in environments where domain knowledge rapidly changes.
Originality/value: This study contributes a novel domain-agnostic framework that combines natural language processing with image analysis to predict concept drift early. This unique approach, which is focused on real-time data sources, offers an effective and scalable solution for addressing the evolving nature of domain-specific concepts in AI applications.

Purpose: Currently, different research conclusions exist about the relationship between relational capital and corporate innovation. The research aims to (1) reveal the actual relationship between executive alumni relations and firm innovation performance, (2) examine the moderating role of executive academic backgrounds, (3) analyze the paths for firms to leverage knowledge spillovers from regional universities to promote firm innovation by their geographic location.
Design/methodology/approach: A social network approach is used to construct alumni relationship networks of A-share listed companies in Shanghai and Shenzhen, China. A two-way fixed effects model is used to assess the impact of firms' structural position in executive alumni networks on firms' innovation performance. In addition, the research also delves into the interactions between knowledge spillovers from geographic locations and executives' alumni networks, aiming to elucidate their combined effects on firms' innovation performance.
Findings: This paper explores the curvilinear relationship between executive alumni networks' centrality and firm innovation within the Chinese context. It also finds that in the positive effect interval on the right side of the “U-shaped,” the industry with the highest number of occurrences is the high-tech industry. Moreover, it elucidates the moderating influence of executives' academic experience on the alumni networks-innovation nexus, offering a nuanced understanding of these dynamics. Lastly, we provide novel insights into optimizing resource allocation to leverage geographic knowledge spillovers for innovation.
Research limitations: The study may not fully represent the broader population of firms, particularly small and medium-sized enterprises (SMEs) or unlisted companies. Future research could expand the sample to include a more diverse range of firms to enhance the generalizability of the findings.
Practical implications: Firstly, companies can give due consideration to the alumni resources of executives in their personnel decisions, but they should pay attention to the rational use of resources. Secondly, universities should actively work with companies to promote knowledge transfer and collaboration.
Originality/value: The findings help clarify the influence mechanism of firms' innovation performance, providing theoretical support and empirical evidence for firms to drive innovation at the executive alumni relationship network level.

Purpose: This study examines the impact of research policy changes on scientific retractions of publications authored by Romanian authors, focusing on national trends and the interplay between policy reforms and publishing practices.
Design/methodology/approach: Using data from the Retraction Watch Database and Web of Science (WoS), 188 unique retractions involving Romanian authors (2000-2022) were analyzed. The study compared retraction patterns before and after the 2016 reforms, which prioritized the publication of articles in WoS-indexed journals over non-WoS outputs.
Findings: The analysis identified two key trends: (1) before the 2016 reforms, retractions predominantly involved non-WoS journals (99 non-WoS retractions to 38 WoS retractions), a trend that reversed post-reform (16 non-WoS to 35 WoS), and (2) while the total number of WoS-indexed retractions increased after the reforms, the retraction rates for WoS articles remained stable. Post-reform reliance on MDPI journals, which have low retraction rates, partially explains this stability. Excluding MDPI publications, retraction rates for articles and reviews increase by 14.91%, aligning with patterns seen elsewhere.
Research limitations: The study focuses on retractions involving Romanian authors, limiting its generalizability. Furthermore, reliance on database records may not fully capture all retractions.
Practical implications: These findings underscore the need for research policy reforms to consider a broader range of effects, and the need for nuanced interpretations of retraction data, which are influenced by a complex range of factors, including specific publisher practices.
Originality/value: This research is the first to investigate the complex relationship between research policy reforms, publisher behavior, and retraction trends.

Purpose: Policies have often, albeit inadvertently, overlooked certain scientific insights, especially in the handling of complex events. This study aims to systematically uncover and evaluate pivotal scientific insights that have been underrepresented in policy documents by leveraging extensive datasets from policy texts and scholarly publications.
Design/methodology/approach: This article introduces a research framework aimed at excavating scientific insights that have been overlooked by policy, encompassing four integral parts: data acquisition and preprocessing, the identification of overlooked content through thematic analysis, the discovery of overlooked content via keyword analysis, and a comprehensive analysis and discussion of the overlooked content. Leveraging this framework, the research conducts an in-depth exploration of the scientific content overlooked by policies during the COVID-19 pandemic.
Findings: During the COVID-19 pandemic, scientific information in four domains was overlooked by policy: psychological state of the populace, environmental issues, the role of computer technology, and public relations. These findings indicate a systematic underrepresentation of important scientific insights in policy.
Research limitations: This study is subject to two key limitations. Firstly, the text analysis method—relying on pre-extracted keywords and thematic structures—may not fully capture the nuanced context and complexity of scientific insights in policy documents. Secondly, the focus on a limited set of case studies restricts the broader applicability of the conclusions across diverse situations.
Practical implications: The study introduces a quantitative framework using text analysis to identify overlooked scientific content in policy, bridging the gap between science and policy. It also highlights overlooked scientific information during COVID-19, promoting more evidence-based and robust policies through improved science-policy integration.
Originality/value: This paper provides new ideas and methods for excavating scientific information that has been overlooked by policy, further deepens the understanding of the interaction between policy and science during the COVID-19 period, and lays the foundation for the more rational use of scientific information in policy-making.

Purpose: The study examines the synergy and hysteresis in the evolution of funding and its supported literature, depicts their temporal correlation mechanism, which aids in improving trend predictions.
Design/methodology/approach: The study uses the LDA model to identify topics in funding texts and supported papers. A cosine similarity algorithm was employed to estimate the nexus between topics and construct the topic evolution time series. Similarly, the hysteresis effect in topic evolution is analyzed based on topic popularity and content, leading to insights into their temporal correlation mechanism.
Findings: The study finds that fund and sponsored paper topics exhibit strong collaboration with a noticeable lag in evolution. The fund topics significantly influence sponsored paper topics after a two-year lag. Moreover, the lag effect is inversely proportional to the topic’s similarity.
Research limitations: We use the LDA model to determine the hysteresis effect in topic evolution despite its limitations in handling long-tail words and domain-specific vocabulary. Furthermore, the timing of the emergence of the focal topic in funds is undermined, affecting the findings.
Practical implications: These findings enhance the accuracy and scientific validity of trend prediction. Estimating and identifying patterns can help technology managers anticipate future research hotspots, supporting informed decision-making and technology management.
Originality/value: This study introduces a research framework to quantitatively and visually analyze the hysteresis effect, revealing the correlation and evolutionary patterns between fund research topics and their funded papers.

The author regrets that the paper titled “Gauging scholars’ acceptance of Open Access journals by examining the relationship between perceived quality and citation impact” (DOI: 10.2478/jdis-2025-0002), as published, contains errors in four of the table captions. For Tables 12–15, “CABS business journals” should read “CABS economics journals.” The tables do have the correct values for the economics journals, and the findings reported in the text do not need revision. The author apologizes for any inconvenience.

Purpose: Scholars face an unprecedented ever increasing demand for acting as reviewers for journals, recruitment and promotion committees, granting agencies, and research assessment agencies. Consequently, journal editors face an ever increasing scarcity of experts willing to act as reviewers. It is not infrequent that reviews diverge, which forces editors to recur to additional reviewers or make a final decision on their own. The purpose of the proposed bibliometric system is to support of editors' accept/reject decisions in such situations.
Design/methodology/approach: We analyse nearly two million 2017 publications and their scholarly impact, measured by normalized citations. Based on theory and previous literature, we extrapolated the publication traits of text, byline, and bibliographic references expected to be associated with future citations. We then fitted a regression model with the outcome variable as the scholarly impact of the publication and the independent variables as the above non-scientific traits, controlling for fixed effects at the journal level.
Findings: Non-scientific factors explained more than 26% of the paper's impact, with slight variation across disciplines. On average, OA articles have a 7% greater impact than non-OA articles. A 1% increase in the number of references was associated with an average increase of 0.27% in impact. Higher-impact articles in the reference list, the number of authors and of countries in the byline, the article length, and the average impact of co-authors' past publications all show a positive association with the article's impact. Female authors, authors from English-speaking countries, and the average age of the article's references show instead a negative association.
Research limitations: The selected non-scientific factors are the only observable and measurable ones to us, but we cannot rule out the presence of significant omitted variables. Using citations as a measure of impact has well-known limitations and overlooks other forms of scholarly influence. Additionally, the large dataset constrained us to one year's global publications, preventing us from capturing and accounting for time effects.
Practical implications: This study provides journal editors with a quantitative model that complements peer reviews, particularly when reviewer evaluations diverge. By incorporating non-scientific factors that significantly predict a paper's future impact, editors can make more informed decisions, reduce reliance on additional reviewers, and improve the efficiency and fairness of the manuscript selection process.
Originality/value: To the best of our knowledge, this study is the first one to specifically address the problem of supporting editors in any field in their decisions on submitted manuscripts with a quantitative model. Previous works have generally investigated the relationship between a few of the above publication traits and their impact or the agreement between peer-review and bibliometric evaluations of publications.

Purpose: This study explores the combined effects of structural and relational embeddedness within alliance networks on firm innovation. By focusing on the interplay between network structures and relationships, this study provides a nonlinear framework to unravel the complex dynamics between alliance networks and firm innovation performance within the manufacturing industry.
Design/methodology/approach: Using social network analysis, this study examines the topological structure of firms’ alliance networks. An exploratory approach involving K-Means clustering and decision tree methods is employed to identify heterogeneous network types within the alliance networks. The analysis further explores the nonlinear relationships between network characteristics, including closeness centrality, betweenness centrality, clustering coefficient, and relational attributes, including collaboration intensity and breadth, and their combined influence on firm innovation.
Findings: The study identified four distinct heterogeneous network types: dyadic, star, ringlike, and complex networks. Each type reveals unique network characteristics and their impact on innovation performance. Key decision rules were extracted, showing that strong relational embeddedness can hinder innovation in dyadic networks, while a greater distance from the central firm correlates with higher innovation performance in star alliance networks. For ringlike alliance networks, moderate cooperation intensity is beneficial for innovation when the clustering coefficient is not high. In complex alliance networks, the combined effects of cooperation intensity, breadth, and clustering coefficient significantly influence innovation.
Research limitations: The research presented in this study, while offering valuable insights into the relationship between alliance networks and firm innovation within the manufacturing sector, is subject to several limitations. A focus on the manufacturing industry may restrict the generalizability of our findings to other sectors, where the dynamics of innovation and collaboration might differ significantly. Additionally, our reliance on patent data, while providing a quantifiable measure of innovation, may overlook other forms of innovation that are equally critical in different contexts, such as service innovations or business model transformations.
Practical implications: This research offers significant insights into how firms can leverage both network structure and relational aspects to enhance innovation outcomes. By revealing the nonlinear and complex interactions between network embeddedness dimensions, this study makes a valuable contribution to both theory and practice. This highlights that strategic management of both structural and relational embeddedness can foster superior innovation performance, offering firms a competitive advantage by optimizing their alliance network configurations.
Originality/value: This study’s originality lies in its examination of the combined effects of structural and relational network embeddings on innovation performance. By identifying distinct network types and their impact on innovation, this study advances the theoretical understanding of how network characteristics interact to shape firm innovation. It contributes to the literature by offering a novel, multidimensional framework that integrates social network theory and resource-based view, providing new insights for firms to leverage their network positions and relationships for competitive advantage.

Purpose: Generally, the scientific comparison has been done with the help of the overall impact of scholars. Although it is very easy to compare scholars, but how can we assess the scientific impact of scholars who have different research careers? It is very obvious, the scholars may gain a high impact if they have more research experience or have spent more time (in terms of research career in a year). Then we cannot compare two scholars who have different research careers. Many bibliometrics indicators address the time-span of scholars. In this series, the h-index sequence and EM/EM’-index sequence have been introduced for assessment and comparison of the scientific impact of scholars. The h-index sequence, EM-index sequence, and EM’-index sequence consider the yearly impact of scholars, and comparison is done by the index value along with their component value. The time-series indicators fail to give a comparative analysis between senior and junior scholars if there is a huge difference in both scholars’ research careers.

Design/methodology/approach: We have proposed the cumulative index calculation method to appraise the scientific impact of scholars till that age and tested it with 89 scholars data.

Findings: The proposed mechanism is implemented and tested on 89 scholars’ publication data, providing a clear difference between the scientific impact of two scholars. This also helps in predicting future prominent scholars based on their research impact.

Research limitations: This study adopts a simplistic approach by assigning equal credit to all authors, regardless of their individual contributions. Further, the potential impact of career breaks on research productivity is not taken into account. These assumptions may limit the generalizability of our findings

Practical implications: The proposed method can be used by respected institutions to compare their scholars impact. Funding agencies can also use it for similar purposes.

Originality/value: This research adds to the existing literature by introducing a novel methodology for comparing the scientific impact of scholars. The outcomes of this research have notable implications for the development of more precise and unbiased research assessment frameworks, enabling a more equitable evaluation of scholarly contributions.

Purpose: Journal Impact Factors and other citation-based indicators are widely used and abused to help select journals to publish in or to estimate the value of a published article. Nevertheless, citation rates primarily reflect scholarly impact rather than other quality dimensions, including societal impact, originality, and rigour. In response to this deficit, Journal Quality Factors (JQFs) are defined and evaluated. These are average quality score estimates given to a journal’s articles by ChatGPT.
Design/methodology/approach: JQFs were compared with Polish, Norwegian and Finnish journal ranks and with journal citation rates for 1,300 journals with 130,000 articles from 2021 in large monodisciplinary journals in the 25 out of 27 Scopus broad fields of research for which it was possible. Outliers were also examined.
Findings: JQFs correlated positively and mostly strongly (median correlation: 0.641) with journal ranks in 24 out of the 25 broad fields examined, indicating a nearly science-wide ability for ChatGPT to estimate journal quality. Journal citation rates had similarly high correlations with national journal ranks, however, so JQFs are not a universally better indicator. An examination of journals with JQFs not matching their journal ranks suggested that abstract styles may affect the result, such as whether the societal contexts of research are mentioned.
Research limitations: Different journal rankings may have given different findings because there is no agreed meaning for journal quality.
Practical implications: The results suggest that JQFs are plausible as journal quality indicators in all fields and may be useful for the (few) research and evaluation contexts where journal quality is an acceptable proxy for article quality, and especially for fields like mathematics for which citations are not strong indicators of quality.
Originality/value: This is the first attempt to estimate academic journal value with a Large Language Model.

Google Gemini 1.5 Flash scores were compared with ChatGPT 4o-mini on evaluations of (a) 51 of the author’s journal articles and (b) up to 200 articles in each of 34 field-based Units of Assessment (UoAs) from the UK Research Excellence Framework (REF) 2021. From (a), the results suggest that Gemini 1.5 Flash, unlike ChatGPT 4o-mini, may work better when fed with a PDF or article full text, rather than just the title and abstract. From (b), Gemini 1.5 Flash seems to be marginally less able to predict an article’s research quality (using a departmental quality proxy indicator) than ChatGPT 4o-mini, although the differences are small, and both have similar disciplinary variations in this ability. Averaging multiple runs of Gemini 1.5 Flash improves the scores.