Search bioRxiv⌕ Search
Home/Archive/bioRxiv

bioRxiv · 10.1101/2020.04.08.031765

Measuring the quality of scientific references in Wikipedia: an analysis of more than 80M citations to over 800,000 scientific articles

Nicholson, J. M.·Uppala, A.·Sieber, M.·Grabitz, P.·Mordaunt, M.·Rife, S.
View original source ↗Export metadata ↓

Abstract

Wikipedia is a widely used online reference work which cites hundreds of thousands of scientific articles across its entries. The quality of these citations has not been previously measured, and such measurements have a bearing on the reliability and quality of the scientific portions of this reference work. Using a novel technique, a massive database of qualitatively described citations, and machine learning algorithms, we analyzed 1,923,575 Wikipedia articles which cited a total of 841,821 scientific articles, and found that most cited articles (58%) are uncited or untested by subsequent studies, while the remainder show a wide variability in contradicting or supporting evidence (2-40%).View Full Text

Source connections

Explore related subjects

scientific communication and educationPreprint

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Nicholson, J. M., Uppala, A., Sieber, M., Grabitz, P., Mordaunt, M., Rife, S.. 2020-04-09. Measuring the quality of scientific references in Wikipedia: an analysis of more than 80M citations to over 800,000 scientific articles. https://doi.org/10.1101/2020.04.08.031765

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related preprints

bioRxiv2026

Evaluating Large Language Models as Tools to Navigate Researchers in Rapidly Evolving Research Landscapes: A Case Study in Cancer Drug Response Prediction

Large Language Models (LLMs) have emerged as promising tools for assisting researchers in automating and accelerating the synthesis of literature reviews. However, their reliability is a significant concern due to issues like factual inaccuracies and hallucinations. The key question is whether LLMs can reliably provide comprehensive, up-to-date overviews and analyses. This study evaluates the performance of three leading LLMs (OpenAI's ChatGPT, Google's Gemini, and DeepSeek) on the complex task of generating a comprehensive survey paper on deep learning for cancer Drug Response Prediction (DRP). By testing both standard and Deep Research (DR) / Deep Think (DT) modes of LLMs with prompts of varying detail, this paper assesses key academic dimensions, including reference management, content quality, and analytical depth. Key findings reveal that while DR modes of LLMs significantly improve reliability by eliminating hallucinations, performance variations exist across models and prompts. A trade-off between reference quantity and integration quality was observed, and even the best-performing models lacked the analytical depth of human experts, often requiring extensive human supervision. The study concludes that LLMs currently serve as powerful assistive tools but still cannot replace the critical validation and synthesis provided by human researchers. Choosing the best LLM to use depends on the task in hand, while several strategies can be implemented to improve the produced output.

scientific communication and education↗
bioRxiv2020

Cooperative efforts on developing vaccines and therapies for COVID-19

Health organizations have always sought partnership to join competencies in innovation, even with fierce competition in this sector. In this pandemic moment it is relevant to observe how organizations behave to seek quick and safe answers. The present research analyzes how the cooperation networks were set off considering the clinical trials on therapies and vaccines that were developed specifically to treat or prevent COVID-19. Social Network Analysis technique was used to build cooperation networks and apply metrics that characterize these connections. There was an evaluation of statistics of Strength of cooperation and Unilateral dependence of cooperation that identify the cooperation strength between two organizations, and the dependence of this relations. A total of 415 clinical trial were identified, of which 42% are in cooperation. From organizations that have partnership, firms are the first, followed by universities. We extracted the main categories that concentrate 74% of partnerships in the trials of antibody, and vaccine. Several organizations cooperate in multiple categories of trials, evidencing the efforts to focus on different strategies to treat the disease. We found high strength of cooperation and an assimetryc dependency between partners, which can be assigned to specialized models of partnership and it occurs in competitive enviroments like this pandemic moment. Cooperation were not limited to geographical proximity and the advent of Chinese players can represent a new change in the biotechnological development axis. Finally, the challenge of finding therapeutic or immunological solutions for COVID-19 demonstrates a clear composition of cooperation groups that complement their skills to manage organizational strategies to beat the pandemic. In this new paradigm, there can be partnerships not only in clinical trial but also in pre-competitive technologies development. This experience is expected to change the way of organizations define their R&D strategies and start to adopt more a collaborative innovation model.

scientific communication and education↗
bioRxiv2020

A Mental Health Paradox: Mental health was both a motivator and barrier to physical activity during the COVID-19 pandemic

The COVID-19 pandemic has impacted the mental health, physical activity, and sedentary behavior of citizens worldwide. Using an online survey with 1669 respondents, we sought to understand why and how by querying about perceived barriers and motivators to physical activity that changed because of the pandemic, and how those changes impacted mental health. Consistent with prior reports, our respondents were less physically active (aerobic activity, -11%, p <0.05; strength-based activity, -30%, p<0.01) and more sedentary (+11%, p<0.01) during the pandemic as compared to 6-months before. The pandemic also increased psychological stress (+22%, p <0.01) and brought on moderate symptoms of anxiety and depression. Respondents whose mental health deteriorated the most were also the ones who were least active (depression r = -.21, p<0.01; anxiety r = -.12, p<0.01). The majority of respondents were unmotivated to exercise because they were too anxious (+8%, p <0.01), lacked social support (+6%, p =<0.01), or had limited access to equipment (+23%, p <0.01) or space (+41%, p <0.01). The respondents who were able to stay active reported feeling less motivated by physical health outcomes such as weight loss (-7%, p<0.01) or strength (-14%, p<0.01) and instead more motivated by mental health outcomes such as anxiety relief (+14%, p <0.01). Coupled with previous work demonstrating a direct relationship between mental health and physical activity, these results highlight the potential protective effect of physical activity on mental health and point to the need for psychological support to overcome perceived barriers so that people can continue to be physically active during stressful times like the pandemic.

scientific communication and education↗