Deceptive Science Maps: Organizational Affiliation and Data Manipulation in Alzheimer's Research

Does corporate authorship increase or decrease deceptive conduct in science? Commercial interests are often viewed as a threat to scientific integrity. Yet when firms build on their science in downstream innovation, fraudulent findings can derail costly development, creating private incentives to produce reliable evidence. We test this logic in Alzheimer’s research, using inappropriate image duplication as an observable marker of deceptive conduct. Combining AI-based detection with manual validation, we document a rising incidence of data manipulation. Such cases are nearly absent in corporate-authored papers and less frequent in academic-industry collaborations, but only when firms exercise meaningful control over the production of evidence. The reduction in problematic duplications is largest when firms are likely to build on the results and in settings where deceptive conduct is more prevalent. However, this difference disappears when research concerns drugs already commercialized by the authoring firm. These findings recast the role of commercial interests in science: they can undermine integrity when legitimating existing products, but strengthen it when firms need solid evidence to build on.