Sort:  

I see common mistakes leaders make all the time:

  • they don't know the error of measurement
  • they do the whole experiment with no clue about how they will analyze the data

The most recent was absolute wasting of time and resources.
While we had 5% error and (maybe?) 2-3% effect, people continued to full themselves that "it's promising" and that "we only need more measurements".
Just change the protocol...

And my favorite, we made the experiment, with animals and we have 4 + 4 + 3, in 3 groups.
Can we do some significance, based on distribution?
Yeah, I can calculate it from my head, without looking the data - not significant.

Coin Marketplace

STEEM 0.28
TRX 0.13
JST 0.032
BTC 60913.71
ETH 2919.21
USDT 1.00
SBD 3.71