Scientific Knowledge Corruption

This EU Horizon 2020 document establishes ethical principles and conduct standards for researchers funded under EU research programs. It outlines obligations around research integrity, data management, and responsible innovation to ensure publicly funded science meets high ethical standards. The code addresses issues like scientific misconduct, transparency, and accountability in European research.

John Ioannidis's landmark 2005 paper demonstrates mathematically that the majority of published research findings are likely false positives, due to low statistical power, publication bias, and researcher degrees of freedom. Using probability modeling, it shows that under common research conditions the post-study probability of a true finding is frequently below 50%. This work catalyzed the modern replication crisis movement across scientific disciplines.

COPE is a global membership organization promoting ethical standards in scholarly publishing, providing guidance, education, and leadership on issues like research integrity, editorial independence, and AI in publishing. With over 14,500 members across 97 countries, it serves as a central authority for publication ethics norms. Its resources include guidelines, position statements, and discussion documents relevant to research integrity challenges.

The Singapore Statement on Research Integrity is the first international effort to establish unified principles and responsibilities for research integrity worldwide, developed at the 2nd World Conference on Research Integrity in 2010. It was produced collaboratively by 340 participants from 51 countries and aims to encourage governments, institutions, and researchers to develop comprehensive standards and codes of conduct promoting honest research globally.

ImageTwin appears to be a service related to image detection or verification, likely for identifying duplicate or manipulated images in scientific publications, but the website is currently in a 'coming soon' state with no substantive content available.

This large-scale study screened over 20,000 papers across 40 scientific journals and found that 3.8% contained problematic figures with inappropriate image duplication, at least half showing signs of deliberate manipulation. The prevalence has risen markedly over the past decade, and journal-level practices like prepublication image screening appear to influence data quality.

For Better Science is an investigative blog by Leonid Schneider that exposes scientific misconduct, data fraud, paper mills, and integrity failures across biomedical and life sciences research. It critically examines problematic publications, institutional cover-ups, and the broader replication crisis. The site serves as a watchdog resource for accountability in academic publishing.

Statcheck is a free online tool that automatically checks statistical results in research papers for inconsistencies and errors, such as mismatches between reported test statistics, degrees of freedom, and p-values. It helps researchers and reviewers identify potential errors in null-hypothesis significance testing (NHST) results. The tool supports efforts to improve scientific integrity and reproducibility.

Crossref Event Data is a service that tracks and aggregates online activity and discussions around scholarly content, collecting data on how research is referenced, shared, and discussed across the web. It provides an open dataset of events linking scholarly works to online sources such as social media, Wikipedia, and news outlets. This helps researchers and institutions understand the broader impact and reach of academic publications.

A preprint study found that AI chatbots like ChatGPT can generate research paper abstracts that are convincing enough to fool scientists into believing they are human-written. The research, posted on bioRxiv in December 2022, demonstrates that current detection methods are unreliable at identifying AI-generated academic content. This finding has sparked debate within the scientific community about the implications for research integrity and the need for better detection tools or policies to address AI-generated submissions.

The Committee on Publication Ethics (COPE) provides comprehensive guidelines for editors, authors, and reviewers on handling research misconduct, fraud detection, and ethical publishing practices. It serves as a central resource for maintaining integrity in academic publishing. The guidance covers issues such as paper mills, plagiarism, data fabrication, and peer review manipulation.

This Nature article examines the problem of paper mills—organizations that produce fraudulent academic papers for sale—and their impact on scientific integrity and the replication crisis. It discusses how systematic fabrication of research undermines trust in published science and proposes strategies for detection and prevention.

Retraction Watch is a blog and database that tracks retractions of scientific papers and other issues of research integrity, providing transparency about errors, fraud, and misconduct in academic publishing. It serves as a critical resource for understanding the scale and nature of problems in scientific literature, including paper mills and reproducibility failures. The site maintains a searchable database of over 45,000 retracted papers.

PubPeer is an online platform enabling post-publication peer review, allowing researchers to comment on and critique published scientific papers anonymously. It has become a major venue for detecting research fraud, data manipulation, image duplication, and paper mill activity. The platform plays a significant role in scientific accountability and the broader replication crisis discourse.

This paper proposes Auto-DSTSGN, an automated dilated spatio-temporal synchronous graph network for traffic prediction in intelligent transportation systems. The key innovation is an automated graph structure search approach that dynamically constructs the spatio-temporal graph to adapt to different data scenarios, rather than using fixed graph construction methods. The model uses dilated layers with increasing dilation factors to capture both short-term and long-term spatio-temporal dependencies more effectively. Experiments on four real-world datasets demonstrate approximately 10% performance improvements over state-of-the-art methods.

The NIH defines research misconduct under Public Health Service policies as fabrication, falsification, or plagiarism in any stage of research, explicitly excluding honest error or differences of opinion. The guidelines establish the three core categories with specific examples, forming the regulatory backbone for research integrity enforcement in federally funded science.

A web-based tool designed to help researchers identify potentially problematic or fraudulent academic papers, supporting research integrity efforts. It screens papers for indicators associated with paper mills, fabricated data, or other forms of scientific misconduct. The tool contributes to combating the replication crisis and improving the reliability of the scientific literature.

The Retraction Watch Database is a searchable repository tracking retractions, expressions of concern, and corrections across scientific literature. It documents the reasons behind retractions—including fraud, data fabrication, and plagiarism—serving as a key resource for assessing scientific integrity. The database supports researchers, journalists, and institutions in monitoring the reliability of published science.

The Retraction Watch Database is a comprehensive, searchable repository tracking retracted scientific papers across disciplines. It provides transparency into the scientific correction process by cataloging retractions, expressions of concern, and corrections with reasons such as fraud, error, or plagiarism. It serves as a critical resource for researchers verifying the integrity of cited literature.