4 ms·

Replication until statistically significance is found?

• ekianjo 7y ago
  In a way yes, because replication lowers the risk that the findings in a single experiment are due to random luck. Anyway there's no way around avoiding replication when you want to prove something.
• buboard 7y ago
  Ehe. Yeah if one insists enough this can happen. It's not a panacea, and really switching to any one measure will cause it to be hacked in the end. The solution is probably multiple objectives, eventually including the reputation of the researchers as well nobody has solved the trust issue