"Did Sally freely φ?" There are two possible questions here that are commonly conflated. On the one hand, we may be asking roughly whether Sally intentionally φ-ed (as opposed to acting out of some unwilled compulsion, temporary insanity, or the like) -- i.e. whether Sally truly made a choice here. This is a question about 'free will' or agency. Alternatively, one might ask a question about 'political freedom', or whether Sally was subject to any kind of coercion or constraint in the options she was choosing between: e.g. did anyone have a gun to her head?
Some philosophers (at least since A.J. Ayer) discussing free will have used the gun case as a paradigm example of 'unfree' choice, for which the coerced agent is not morally responsible. This strikes me as simply mistaken. Sally is still responsible for her choice, within the given constraints. If the bandit tells her to kick me in the shins or he'll shoot, then Sally has a choice: she can choose between kicking me in the shins or getting shot. So she kicks me. Good for her, she made the right choice. This is worth emphasizing: it's not that she wasn't responsible for the choice -- on the contrary, she retained her free will, and moreover, she exercised it correctly (given the constraints). She did exactly what she had most reason to do.
Compare this to a case where Sally falls under a compulsion to kick me in the shins (maybe she was hypnotized). Notice how different the moral implications are. Here the action is wrong -- there's no justification for kicking me in the shins -- but Sally isn't responsible for it.
The question of coercion may be relevant when we want to clarify what reasons Sally had for acting as she did, or what options she was choosing between. But that is not a question relevant to the free will debate. There we are instead concerned with the question whether it was really Sally choosing at all. And to answer that question in the negative requires compulsion, not mere coercion.