An interesting study on how warfare affects birds in #Ukraine. The authors combined open-source intelligence, field observations, and eyewitness reports to show that even highly mobile species often remain in combat-affected areas: they die from explosions, lose nesting sites, and face the destruction of habitats and changes in water systems.

:doi: https://doi.org/10.1080/24750263.2025.2531125

In many ways, this is one of the first attempts to systematically document the real-time impact of #war on #biodiversity.

Our new paper asks a simple question: What happens to review articles in the age of generative #AI?

📄 https://doi.org/10.1002/leap.2045

Our answer: #genAI does not make reviews obsolete – it exposes which formats no longer work.

Descriptive, template-driven reviews are easy to scale and lose their signalling value. Reflexive, agenda-setting reviews still matter because judgment and interpretation cannot be automated.

#Publishing #GenerativeAI #ReviewArticles #PeerReview #ResearchInfrastructure

Can we trust “guidelines on how to write a scientific paper”? We analysed 71 “Write a Scientific Paper” guidelines that were widely used and cited for years as best practices.

:doi: https://doi.org/10.1080/10875301.2026.2622096

Facts:
▪️ 555 citations
▪️ 48 papers carry an editorial expression of concern
▪️ 30 (42%) have been retracted!

#AcademicPublishing #ResearchIntegrity #Retractions #ScholarlyCommunication #Bibliometrics #PublicationEthics #PeerReview

Today I completed my author’s course for early-career researchers and PhD students at our university.

We focused on things that are often missing from formal curricula: how to design a study that does not fall apart halfway, how to work with scholarly sources, and how to keep coherence between research questions, methods, and results.

I hope it became one small step toward a more conscious and sustainable research path for the participants.

#PhDlife #Research #AcademicSkills #UniversityLife

Sitting without electricity and heating, but still thinking about… open #bibliometric data. Sharing our presentation from Bergen 2025. Even in these conditions, we keep building resilient research infrastructures:

👉 https://doi.org/10.6084/m9.figshare.30753224

The bibliometrics market is a textbook case of market failure: monopolies dominate, national research stays invisible, and profit beats #data quality. That’s why national infrastructures and #openmetadata really matter.

#OpenScience #OpenData #SciencePolicy

A new #Science review of 34,000 world-class performers shows a surprising pattern: most top scientists, athletes, musicians, and chess players were not child prodigies. About 90% of elite adults were different people than the early “stars.”

📄 https://doi.org/10.1126/science.adt7790

Future champions often developed slower, explored multiple disciplines, and avoided early specialization. Early over-training may help kids win, but it rarely creates world leaders. 🤔

#Education #Creativity #Learning #Metascience

@tobychev Agreed - the novelty is less that bias exists and more how large it is under controlled conditions. The blinding point is important: a two-stage design (blind first, unblind later) would be a great next step to separate analytic flexibility from expectation effects.

A new experiment asked 158 researchers in 71 teams to analyse the same #data and answer the same question: does immigration affect support for welfare policies? The results were all over the place - from strongly negative to strongly positive:

:oa: https://doi.org/10.1126/sciadv.adz7173

The key finding: researchers’ political views mattered. Pro-immigration teams were more likely to find positive effects, while anti-immigration teams found negative ones.

#Ideology #OpenScience #Replication #Metascience

@g3om4c @mike @tillgrallert @hvdsomp @paulwalk This is a really timely piece of work! 👏 It puts hard numbers on a problem that many people in the univ repository and #openscience communities sense intuitively but rarely see measured at scale. The ideas of dead and zombie repositories are especially useful + the evidence on dead on arrival references is striking. 😱 So, this preprint does an excellent job of showing how fragile, supposedly stable scholarly infrastructure can be. 😥

Our new paper in Scientometrics with @mikaellaakso & Zehra Taşkın: https://doi.org/10.1007/s11192-025-05535-4

We map 19,000+ university journals worldwide and show a major blind spot in research evaluation: most are invisible in selective indexes. UJs sustain bibliodiversity, multilingualism, SSH, and Diamond OA — yet remain structurally undervalued.

#ScholarlyCommunication #Bibliodiversity #DiamondOA #UniversityPublishing

@thea @tillgrallert @g3om4c Exactly - and it nicely shows that this isn’t about grand reforms, but about routine maintenance done well. 👏

@thea Thanks for sharing, that aligns very closely with what we’re seeing elsewhere. OAI-PMH keeps resurfacing as a basic but fragile dependency for discovery and reuse.

@g3om4c @mike @tillgrallert @hvdsomp @paulwalk I totally agree. Even partial empirical evidence already helps shift the discussion from assumptions to responsibility. Whether it "shames" 😳 institutions into action is uncertain, but it does make denial harder.

@nyhan @hvdsomp This really captures the problem. From the outside, it looks like a small technical fix, from the inside, it’s often entangled with governance, priorities, and ownership. That opacity is what makes advocacy so hard.

@mike @g3om4c @tillgrallert @hvdsomp @paulwalk It’s encouraging to see these concerns translated into concrete analysis. I’ll be very interested to read it once it’s out.

@g3om4c That sounds painfully familiar. Launching infrastructure is often treated as a project decision, while persistence and URI stewardship are in fact long-term institutional commitments, and that mismatch tends to surface only years later.

@hvdsomp I agree, and the irony is that when technology fails, it’s usually treated as an exception rather than as the core of the problem.

@arjen That’s a good point. Search providers often become the first to expose infrastructural fragility, because broken protocols only show up under real operational use.

@tillgrallert Thank you. Agreed, this is a very concrete infrastructural issue, and @NFDI seems like exactly the right place to have that discussion.

The study shows that around half of supposedly "open" repositories are not technically open at all - their OAI-PMH endpoints are broken, outdated, or misconfigured:

:doi: https://doi.org/10.1177/01655515251396902

For me, this paper is not a revelation, but a confirmation of a simple point: open science does not start with policies or declarations. It starts with routine infrastructural work - making sure that the basic protocols actually work.

#OpenScience #FAIRdata #Repositories #OAIPMH