The second one would be my answer but also an answer that is not up there is that during an experiment when repeating the trials you might have to set up a new hypothesis every time and have to review the data several different times because of different results from each trial
