We seem to have opposite approaches to when hypotheses are formed—look at data, then form hypotheses vs. understand SOTA to form a hypothesis before looking at

We seem to have opposite approaches to when hypotheses are formed—look at data, then form hypotheses vs. understand SOTA to form a hypothesis before looking at the data.