Hacker News Reader

What is the proposed mechanism for Mars getting the outcomes to be positive?

I understand the implied incentives, but either there is fraud in the paper, or the methodology should reveal how the scientists contrived false outcomes.

Otherwise, why does the paper conclude that milk chocolate - Mars' biggest source of revenue from chocolate - has no association with good outcomes and has association with weight gain?

View on HN

P hacking is easy. Get test and control groups. Measure 40 health attributes. Give test group your special compound. Measure 40 health attributes again. Keep only results that show positive benefit (avg. around 2 by random chance). Publish, profit. Data would be clean and does not need to be manipulated. The overall scientific process is rife to be corrupted as long as tester and measurer are the same party, especially when there is financial incentive.

Not getting the next grant is hurting your income. So obvious.

See second paragraph.

Reply on news.ycombinator.com