Addressing Researcher Fraud: Retrospective, Real-Time, and Preventive Strategies — Including Legal Points and Data Management That Prevents Fraud

What if the biggest threat to science isn't bad hypotheses but researchers faking their data? Kennedy tackles this head-on, drawing from decades of hands-on experience -- including running an actual sting operation back in the 1970s and spending 15 years in FDA-regulated drug trials where data integrity is literally a legal requirement. He lays out three ways to fight research fraud. First, the usual approach: investigating after someone blows the whistle. Problem is, by then the evidence is murky, standards are loose, and you often can't even prove who did it. Second, catching cheaters in the act with sting operations -- beautifully conclusive but almost impossible to pull off in practice. Third, and this is the real gem: preventive measures borrowed from the pharmaceutical world that most academic researchers have simply never adopted. Kennedy spells out eight specific practices -- things like keeping tamper-proof audit trails, locking down who can touch the raw data, validating your software, running independent audits, and making datasets public. He even audited a flagship parapsychology study (the Transparent Psi Project) and found a significant software bug lurking in what was supposed to be a gold-standard registered study. That's a wake-up call. He proposes a new badge system so readers can instantly see which studies actually meet these standards. And here's the kicker for the future: AI is about to make fraud dramatically harder to detect, raising the stakes for getting these safeguards in place now.