It should also be noted that before collecting any data, one ought to choose the significance level. This is done for several reasons. First and foremost, it is done to prevent the data quantity from influencing one’s decision. Second, the same applies for the quality. In short, in order to ascertain that bias didn’t enter such a study, strict order should be followed.
What will happen when you allow the data to dictate your decision? Well, for one, it will be very similar to a roulette table, where you are about to lay your bets after the ball has just landed on a particular number. That as you surely understand would constitute bias.
What else would be wrong to not follow this order?