Weekly digest: open data, machine learning and social media

Mark Elms

This week, we highlight an upcoming webinar on the future of open data, as well as a virtual workshop on machine learning for research assessment. We listen to a podcast discussing the use of social media in medical communications. We read about how open access should be normalized and about scientific misconduct in the form of questionable error bars and image manipulation. Finally, we learn about new software to potentially root out paper mills and about how the COVID-19 pandemic affected journal impact factors.

To engage with:

The past, present and future of open data via Digital Science

This year, figshare celebrated its 10th anniversary of providing an online open access repository for researchers, pharma companies, publishers and others, allowing them to freely share data, research and papers. As 2022 comes to an end, this webinar, hosted by Digital Science on 13 December 2022, will discuss with figshare the current open data landscape, the recent rise of national and international open access mandates, and what the next 10 years could bring for figshare and the whole open data community.

Rise of the machines via Eventbrite

The scientific research world often uses quantitative indicators and metrics to assess the impact of their research outputs. Recently, machine learning and artificial intelligence techniques have been adopted to automate the assessment of research. On 12 December 2022, the Research on Research Institute will host a free, virtual workshop to discuss the potential advantages and pitfalls of adopting these techniques, especially in regards to their impact on responsible research assessment, as outlined in the Research Excellence Framework and the Future Research Assessment Programme. You can reserve a spot at the webinar here.

To listen to:

Using social media in medical communications via The Publication Plan | 16-minute listen

In the latest International Society for Medical Publication Professionals (ISMPP) InformED podcast series, Jennifer Ghith (Senior Director, Omnichannel Strategy and Innovations Lead at Pfizer) discusses the opportunities, challenges and solutions that social media can bring to medical communications. In the first part of this two-part series, Jennifer explores social media plans, creating content and understanding your audience – while remaining code compliant. You can listen to the second part of the series here.

To read:

Open access should no longer be exciting; it should be the norm via cOAlition S | 6-minute read

Open access should no longer be considered new but rather the norm for scientific publishing. This is the opinion of Sally Rumsey (formerly Jisc’s open access expert) who, until July 2022, spent most of her time supporting the Plan S initiative. Sally argues that, having been around for over 20 years and in light of recent events such as the COVID-19 pandemic, open access publishing should now be considered the default setting and that restricted access publications should be highlighted, rather than the other way around.

Spilling the T on scientific misconduct via Retraction Watch | 3-minute read

Scientific misconduct takes many forms and may be far more prevalent than previously thought. But this recent case might be one of the most egregious examples. In a July 2022 article published in the Hindawi journal Advances in Materials Science and Engineering, the error bars in a figure appear to simply be capital letter Ts, which have been aligned with the tops of the bars in a bar chart. Further examination revealed other questionable practices, including a data availability statement affirming that “Data sharing is not applicable … as no data sets were generated or analysed during the current study”, which is a surprisingly honest and transparent assessment of their results.

The frequency of image manipulation via STAT | 6-minute read

Image manipulation, especially since the advent of graphic design software such as Adobe Photoshop, is not rare in scientific publications. According to Retraction Watch, a paper retraction due to image manipulation happens once every other day, with high-profile researchers often at the centre of occurrences. For instance, only last month, Stanford University opened up investigations into its own president over allegations of image manipulation. Image manipulation is a serious problem as it can result in researchers trying to build on work that may be entirely fictional. Also, image manipulation is frequently not caught by the publisher or during peer review but by unpaid sleuths after publication. Subsequent retractions and corrections can take years to come about, if at all.

An online integrity hub to root out paper mills via Nature | 5-minute read

Paper mills are an increasingly recognized problem in the publishing industry, and publishers are now trialling an automatic system to help flag submitted manuscripts that have paper mill hallmarks. This system is an online integrity hub, developed by STM alongside 24 publishers and scholarly analytics providers, which aims to root out paper mill submissions, as well as searching for other instances of scientific misconduct, such as image manipulation and data fabrication. The prototype integrity hub tools are currently being trialled by various publishers, and STM hopes to make a version of the integrity hub available for wider use by early 2023.

The effect of the COVID-19 pandemic on journal impact factors via Journal of Medical Internet Research Preprints | 18-minute read

The effects of the COVID-19 pandemic were global, affecting practically every corner of the world and every industry. This includes the publishing industry, as this article published as a preprint in the Journal of Medical Internet Research explains. The article looked at how the COVID-19 pandemic affected the journal impact factors of six high-impact medicine journals – Annals, the British Medical Journal, The Lancet, Nature, the New England Journal of Medicine and the Journal of the American Medical Association. The researchers found that, during the pandemic, impact factors, publication rates and citation rates were positively skewed by COVID-19 publications published across the six journals. They interpreted these results by saying that this could “cast doubt on the reliability of highly susceptible impact factors” and could result in researchers exploiting this effect by “monopolising their research on COVID-19”.

Have you watched our Open Pharma Symposium ‘Who can we trust? Open science and pharma research’? Watch it here on our YouTube channel!