This article makes a good case why whistleblowing is such a difficult thing to do. In a broader context, though, how much of the irreproducibility problem (estimated around 60-70%) can be attributed to outright fraud? Only about a percent or less of the literature is retracted due to fraud, which means that every other person is engaged in deliberate misconduct, which I find hard to believe, or that these whistleblowers are risking huge personal backlash to correct only a few percent of the literature, leaving the majority of irreproducible work untouched.
There must be a better way...
I think focusing on a way to shift the incentives so that doing careful reproducible work is rewarded will yield much better dividends, though misconduct won't ever go away, we can make it better for people who are doing the right thing, instead of focusing on punishing those who do wrong. Here's one approach I've been involved with: https://www.scienceexchange.co...