Search bioRxivSearch

bioRxiv · 10.1101/2021.02.02.429450

The effect of mentee and mentor gender on scientific productivity of applicants for NIH training fellowships

Abstract

Several studies have suggested that women in science are less productive than men, and that this gap contributes to their under-representation in the ranks of senior researchers. However, few studies have examined the role of mentoring, and in particular mentor gender, on the productivity of female scientists early in their careers. Such efforts are limited by the difficulties of unambiguously linking mentees to their mentors and measuring the research productivity resulting from those relationships. Here we use our novel author disambiguation solution to investigate the role of self-identified gender in mentorship of 12,932 trainees who either successfully or unsuccessfully applied to the National Institutes of Health for research fellowships between fiscal years 2011 and 2017, applying a multi-dimensional framework to assess productivity. We found that, after normalizing for the funding level of mentors, the productivity of female and male mentees is indistinguishable; it is also independent of the gender of the mentor, other than in measures of clinical impact, where women mentored by women outperform other mentee-mentor dyads.

Explore related subjects

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Yu, H., Willis, K. A., Litovitz, A., Harriman, R. M., Davis, M. T., Meyer, P., Busse, B., Meseroll, R. A., Wijayatilake, H. D., Perkins, M. J., Anderson, J. M., Santangelo, G. M.. 2021-02-03. The effect of mentee and mentor gender on scientific productivity of applicants for NIH training fellowships. https://doi.org/10.1101/2021.02.02.429450

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

KEEP EXPLORING

Related preprints

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

Applying And Promoting Open Science In Ecology - Surveyed Drivers And Challenges

Open Science (OS) comprises a variety of practices and principles that are broadly intended to improve the quality and transparency of research, and the concept is gaining traction. Since OS has multiple facets and still lacks a unifying definition, it may be interpreted quite differently among practitioners. Moreover, successfully implementing OS broadly throughout science requires a better understanding of the conditions that facilitate or hinder OS engagement, and in particular, how practitioners learn OS in the first place. We addressed these issues by surveying OS practitioners that attended a workshop hosted by the Living Norway Ecological Data Network in 2020. The survey contained scaled-response and open-ended questions, allowing for a mixed-methods approach. Out of 128 registered participants we obtained survey responses from 60 individuals. Responses indicated usage and sharing of data and code, as well as open access publications, as the OS aspects most frequently engaged with. Men and those affiliated with academic institutions reported more frequent engagement with OS than women and those with other affiliations. When it came to learning OS practices, only a minority of respondents reported having encountered OS in their own formal education. Consistent with this, a majority of respondents viewed OS as less important in their teaching than in their research and supervision. Even so, many of the respondents suggestions for what would help or hinder individual OS engagement included more knowledge, guidelines, resource availability and social and structural support; indicating that formal instruction can facilitate individual OS engagement. We suggest that the time is ripe to incorporate OS in teaching and learning, as this can yield substantial benefits to OS practitioners, student learning, and ultimately, the objectives advanced by the OS movement.

scientific communication and education

An exploratory analysis of 4844 withdrawn articles and their retraction notes.

The objective of our study was to obtain an updated image of the dynamic of retractions and retraction notes, retraction reasons for questionable research and publication practices, countries producing retracted articles, and the scientific impact of retractions by studying 4844 PubMed indexed retracted articles published between 2009 and 2020 and their retraction notes. RESULTSMistakes/inconsistent data account for 32% of total retractions, followed by images(22,5%), plagiarism(13,7%) and overlap(11,5%). Thirty countries account for 94,79% of 4844 retractions. Top five are: China(32,78%), United States(18,84%), India(7,25%), Japan(4,37%) and Italy(3,75%). The total citations number for all articles is 140810(Google Scholar), 96000(Dimensions). Average exposure time(ET) is 28,89 months. Largest ET is for image retractions(49,3 months), lowest ET is for editorial errors(11,2 months). The impact of retracted research is higher for Spain, Sweden, United Kingdom, United States, and other nine countries and lower for Pakistan, Turkey, Malaysia, and other six countries, including China. CONCLUSIONSMistakes and data inconsistencies represent the main retraction reason; images and ethical issues show a growing trend, while plagiarism and overlap still represent a significant problem. There is a steady increase in QRP and QPP article withdrawals. Retraction of articles seems to be a technology-dependent process. The number of citations of retracted articles shows a high impact of papers published by authors from certain countries. The number of retracted articles per country does not always accurately reflect the scientific impact of QRP/QPP articles. The country distribution of retraction reasons shows structural problems in the organization and quality control of scientific research, which have different images depending on geographical location, economic development, and cultural model.

scientific communication and education