I disagree, if you have a vested interested you're likely to cheat, not every kind of cheating is visible in a paper.

If you e.g. pump out 100 studies and then make a perfect reasoning on the 1 study which had significant results in your favor, then what you did wrong can't be seen in the one paper you published.

As long as the studies aren't preregistered, you have to discount those with vested interests.