“Trust but verify” is a good principle, as is “defense in depth”. It’s very easy to forge numerical data, to take one example, but it’s still worth trying to catch those forgeries (eg via Benford’s Law), b/c anything that reduces the fraud rate also reduces the load on trust as a supporting wall

🏆 Institutional: The Brazilian Reproducibility Initiative is a nationwide effort to evaluate research results in laboratory biology & the largest coordinated replication effort in the field worldwide, showcasing the potential of country-level research improvements. @redebrrepro.bsky.social (3/5)

🏆 Individual: @simine.com, psychologist at @unimelb.bsky.social & editor-in-chief of Psychological Science, is recognized for pioneering methodological rigor, reproducibility & collaborative research, driving initiatives such as @improvingpsych.org & the journal Collabra @ucpress.bsky.social. (2/5)

I have been warning for this for years.

2. Annie Whamond @anniewham.bsky.social - Analysing discussion text and references of paper mill articles in high-impact factor cancer journals.
Not all paper mill articles are easy to recognize; they are published in medium/high IF journals, authors are related to t the topic.
#AIMOS2025

AW: Set of papers where the discussion sounded like a second introduction (background information) with many new references, much more than other control papers.
Likely citation manipulation. This could help sleuths find more of such papers.
* But be aware about ESL language issues.
#AIMOS2025

3. Luciana Machado @machadolac.bsky.social (online): A New Global Alliance Tackling Research Integrity Challenges.
Many great initiatives deal with science integrity, but coordination is important. We connect organizations.
Reach: interactive digital magazine.
www.sci-integrity.com
#AIMOS2025

LB: Meta-research is like a pug wearing a pug-suit. We are doing research on research (less cute though!)
We looked at 25 meta-research studies including 2923 studies. Studies with statistically significant efficacy results are 30% higher among industry-sponsored studies than non-industry
#AIMOS2025

LB: Bias can occur at many stages of research. We can study these e.g. through documents obtained in legal requests.
Industry agendas are not aligned with public health questions or prevention.
www.tandfonline.com/doi/full/10....
ajph.aphapublications.org/doi/abs/10.2...
#AIMOS2025

LB: Obtaining industry documents (through e.g. a court case), we found that negative results (non-significant) were not published. Selective publishing.
It's not just drugs, tobacco, food, but also PFAS (forever chemicals.
www.nejm.org/doi/full/10....
#AIMOS2025