Frontiers in research metrics and analytics最新文献

Background: Ready-to-eat (RTE) foods are increasingly consumed worldwide due to urbanization, dietary shifts, and globalization of food systems, yet they remain a significant vehicle for foodborne diseases. Despite the growing body of research, a systematic mapping of global scientific trends in this area has been lacking.

Methods: We conducted a bibliometric analysis of RTE food microbiology research indexed in Scopus from 1973 to 2025. Publication trends, citation patterns, leading authors, institutions, journals, and country-level contributions were assessed using Bibliometrix. Network analyses (co-authorship, bibliographic coupling, and keyword co-occurrence) were performed using VOSviewer to identify collaborative structures and thematic evolution.

Results: A total of 780 publications across 256 sources were identified, growing at an average annual rate of 6.9%. The field accumulated 19,811 citations, with highly cited works between 1996 and 2015 establishing its intellectual foundations. China, Italy, and the USA led in productivity, while Public Health England, Universidad de Córdoba, and China's Institute of Microbiology were major institutional hubs. Journal of Food Protection, International Journal of Food Microbiology, and Food Control dominated as key sources. Collaboration networks revealed strong international linkages, particularly among high-income countries. Keyword analysis showed two thematic axes: persistent focus on classical pathogens (Listeria monocytogenes, Salmonella, Bacillus cereus) and emerging concerns related to convenience foods, and fast foods.

Conclusion: Research on the microbiological safety of RTE foods has grown steadily, reflecting global recognition of its public health significance. However, outputs remain concentrated in high-income countries, while low- and middle-income regions with high foodborne disease burdens are underrepresented. Future research should prioritize equitable global participation, integration of genomic and omics tools, and translation of findings into food safety policy. This bibliometric evidence highlights the need for stronger international collaboration to ensure the microbiological safety of RTE foods in an era of rapid dietary transition.

Introduction: The rapid integration of artificial intelligence (AI) in scientific research has introduced new challenges to academic integrity, with increasing concerns about AI-related article retractions. This study conducts a comprehensive bibliometric analysis of retracted AI-related articles to characterize their prevalence, causes, and impact on scholarly communication.

Methods: A systematic search was performed in Scopus using the terms "Artificial Intelligence" OR "AI" AND "retract^*" without restrictions on publication year or language. Bibliometric parameters including publication timelines, journal metrics, citation counts, and retraction characteristics were analyzed using VOS Viewer, Bibliometrix, and SPSS. Statistical tests assessed correlations between key variables.

Results: From an initial yield of 1,152 articles, 335 retracted publications met inclusion criteria after duplicate removal and screening. The analysis revealed that 46.3% (155/335) of retractions occurred in 2023, with a median retraction time of 550 days post-publication. Engineering accounted for 30.4% (102/335) of retractions, while 72.2% (243/335) originated from China. Compromised peer review was the most common retraction reason, though 37.9% (127/335) lacked specific justification. Strikingly, 51.1% (172/335) of retracted articles-maintained field citation ratios >1, indicating persistent scholarly influence. Articles in special issues showed significantly faster submission-to-acceptance timelines (p = 0.016). Journal editors initiated 98.5% (330/335) of retractions, while author responses revealed disagreement in 35.4% (34/96) of cases where feedback was available.

Discussion: This study highlights systemic vulnerabilities in AI-related research publication, particularly concerning peer review integrity and prolonged retraction timelines. The continued citation of retracted articles underscores the need for improved retraction alert systems. These findings call for stronger ethical guidelines and technological safeguards to maintain trust in AI-driven scholarly outputs.

Open Science aims to make research more transparent, reusable, and socially valuable, yet adoption may lag where assessment emphasizes journal prestige over openness. This study examines how research-assessment incentives align with Open Science practices in Ecuador and identifies policy levers associated with change. Using a mixed-methods design, we combine a review of national and institutional policies, a bibliometric analysis of Ecuador-affiliated outputs from 2013-2023, and a nationwide researcher survey (n ≈ 418), analyzed with multilevel logistic models, multinomial logit, and negative binomial regressions. Scientific output increased markedly, peaking at 5,070 articles in 2023; 66.7% were open access, predominantly via gold routes. In 2021, 59.3% of citations were self-citations. Despite high familiarity with Open Science (85%), implementation was limited: 22% reported depositing data and 35% publishing via diamond or gold routes. Greater reliance on journal-centric metrics was associated with lower odds of adopting open practices (odds ratio ≈ 0.72), while comprehensive institutional support-repositories with deposit mandates, research-data services, and licensing guidance-was associated with higher odds (odds ratio ≈ 1.65). Sensitivity to article processing charges was associated with shifts toward green and diamond routes. Findings suggest that socio-institutional factors dominate barriers and that aligning rules, services, and responsible assessment may help make openness the default, improving quality, equity, and reuse.

Narrative research is intended to explore human experiences. However, there are ethical dilemmas that challenge researchers beyond formal protocols. This review examines 16 empirical studies (2014-2023) alongside insights from Eastern wisdom traditions, drawing on the experiences of three university faculty members who have employed narrative inquiry methodology in their graduate-level research to explore ethical dilemmas and shortcomings. This review identifies key recurring dilemmas in narrative research, including navigating informed consent, ensuring anonymity/confidentiality, managing power dynamics, mitigating emotional vulnerability, and respecting cultural sensitivity. The findings feature ethical integrity that relies on continuous reflexivity, relational ethics, and trust-building-principles reflected in Eastern concepts such as Dharma (i.e., righteous duties) and Karma (i.e., selfless actions). The study emphasizes the importance of context-sensitive ethical practices that prioritize participant dignity and the researcher's integrity. The article addresses the implications of creating ethical dilemmas in narrative research guidelines, provides ongoing ethical training, and promotes collaborative learning among researchers to enhance the trustworthiness of qualitative research in general and narrative research in particular.

Altmetrics have emerged as a complementary tool to traditional citation-based metrics in the assessment of scholarly impact. Unlike traditional metrics that primarily capture academic citations over long periods, altmetrics reflect immediate online attention across platforms such as Twitter, blogs, news outlets, and Mendeley. This article critically examines whether altmetrics can serve as a substitute for traditional metrics by exploring their strengths, limitations, disciplinary variations, and correlation with conventional indicators. Through a review of recent empirical studies and theoretical debates, the article argues that while altmetrics offer valuable insights into social impact and engagement, they are not yet mature or standardized enough to fully replace traditional metrics. Instead, a hybrid model that integrates both systems may offer a more holistic and inclusive measure of research influence.

Introduction: Open innovation has become a central mechanism for enhancing university-industry collaboration (UIC), fostering the co-development of innovative and socially responsive solutions. As organizations increasingly embrace openness and knowledge-sharing practices, understanding the evolution of open innovation in university-industry collaboration (OIUIC) is critical amid accelerating digitalization and mounting sustainability imperatives.

Methods: This review maps the conceptual structure of OIUIC research from 2003 to 2024 by applying co-word analysis and social network mapping to a dataset of 2,601 articles indexed in Scopus. We extracted and standardized 5,269 unique keywords, constructed co-word networks to identify thematic clusters, and deployed network metrics to reveal patterns of scholarly collaboration and influence.

Results: The analysis uncovered five dominant keyword clusters: "technology transfer," "university-industry knowledge transfer (UIKT)," "knowledge transfer," "academic entrepreneurship," and "university," which collectively define the field's conceptual architecture. Geographically, the United Kingdom leads in publication output, while Research Policy and The Journal of Technology Transfer emerge, respectively, as the most cited and the most prolific journals. Network metrics further highlight key author and institution hubs that bridge thematic communities.

Discussion: By synthesizing major themes and research clusters, this review provides a comprehensive overview of the OIUIC intellectual landscape. Our findings offer critical insights for researchers and policymakers, suggesting priority areas for future inquiry, such as digital transformation, sustainability integration and cross-regional partnership models, and informing evidence-based policy development to strengthen inclusive and adaptive innovation ecosystems.

There is concern about the levels of stress faced by college students and their effects on mental health and academic performance. This study aimed to characterize academic stress levels in college students, using data mining algorithms to classify and predict risk patterns. Data were collected from 287 students using the SISCO Academic Stress Inventory, and classification algorithms and association rules were applied using WEKA software. The results revealed that 75.3% of the students experienced high stress levels, primarily linked to psychological reactions and academic demands. It also compared the predictive performance of 13 algorithms, where J48, LMT, and SimpleLogistic achieved classification accuracies above 89%, surpassing results previously reported in similar educational contexts. Association rule mining further showed that being single and childless was strongly correlated with elevated stress levels, highlighting demographic risk profiles often overlooked in earlier research. By integrating predictive modeling with demographic and behavioral factors, this study extended prior literature by showing how data mining can simultaneously classify and explain academic stress, offering actionable insights for universities to design targeted, evidence-based interventions.

This scoping review deals with a major trend in qualitative data analysis: thematic analysis (TA) that provides a general framework to develop relatively transparent processes; TA thus helps mitigate long-standing concerns with allegedly subjective aspects of qualitative research. The review examines articles published in the top-ranked academic journals in the research area "communication" (n = 342). It illustrates that TA has quickly become more popular over recent years, complementing longer established qualitative methods. The analysis also reveals that TA is a flexible tool that has been successfully applied to make sense of a wide array of qualitative data. Building on these findings, we deduce practical advice for researchers applying TA and make suggestions on how to improve on current TA practices by, first, documenting common features of successful TA articles (best practice), and, second, identifying apparent superficialities and untapped potentials.

Background: Manual quality assessment of systematic reviews is labor-intensive, time-consuming, and subject to reviewer bias. With recent advances in large language models (LLMs), it is important to evaluate their reliability and efficiency as potential replacements for human reviewers.

Aim: This study assessed whether generative AI models can substitute for manual reviewers in literature quality assessment by examining rating consistency, time efficiency, and discriminatory performance across four established appraisal tools.

Methods: Ninety-one systematic reviews were evaluated using AMSTAR 2, CASP, PEDro, and RoB 2 by both human reviewers and two LLMs (ChatGPT-4.0 and DeepSeek R1). Entropy-based indicators quantified rating consistency, while Spearman correlations, receiver operating characteristic (ROC) analysis, and processing-time comparisons were used to assess the relationship between time variability and scoring reliability.

Results: The two LLMs demonstrated high consistency with human ratings (mean entropy = 0.42), with particularly strong alignment for PEDro (0.17) and CASP (0.25). Average processing time per article was markedly shorter for LLMs (33.09 s) compared with human reviewers (1,582.50 s), representing a 47.80-fold increase in efficiency. Spearman correlation analysis showed a statistically significant positive association between processing-time variability and rating entropy (ρ = 0.24, p = 0.026), indicating that greater time variability was associated with lower consistency. ROC analysis further showed that processing-time variability moderately predicted moderate-to-low consistency (AUC = 0.65, p = 0.045), with 46.00 seconds identified as the optimal cutoff threshold.

Conclusion: LLMs markedly reduce appraisal time while maintaining acceptable rating consistency in literature quality assessment. Although human validation is recommended for cases with high processing-time variability (>46.00 s), generative AI represents a promising approach for standardized, efficient, and scalable quality appraisal in evidence synthesis.