An interesting issue. One way around it would be to enact the law in two states that have a similar environments. Continuing the example with gun control if there were two states with similar situations, both of them have an abnormally high gun crime rate, one of them could serve as a control state. If say 10 pairs of states were compared it would serve as a reasonable experiment on whether the law should be passed nation wide or not. Even better if evidence comes up that passing the law would only be beneficial in certain circumstances then it would help reveal the underlying causes of the problem in the first place.