The bell inequalities are a concrete evidence to dismiss 1. Say we flip a coin (just for the sake of the argument - superposition only works for small things),

The bell inequalities are a concrete evidence to dismiss 1. Say we flip a coin (just for the sake of the argument - superposition only works for small things), and we're arguing whether the coin is really heads or really tails before we look at it, or whether it's in a superposition. The bell inequalities give us a prediction on the distribution of results we should see if 1. was true or if 2. was true. Turns out the distributions we see in experiments are consistent with 2 but not with 1. The measured values can not be true if the coin is either head or tails. Of course, there are subtleties involved, but that's one of the main evidences we have.