Peer validation is not a get out of gaol free card. In practice, it is flawed. Then there is reproducibility, but when you talk about psychology, for example, it is never going to have the rigour of a subject like physics. In fact, it is a field that is notorious for people fiddling with their own data and getting away with it.
The best thing I have heard as a real explanation for what the "scientific method" is, at its core, is parsimony. As in, a process of creating a simplified model of natural phenomena that is neither too simple nor too complex to be useful. But today it is accepted that many things which are not useful at all, overly simple, or overly complex are all "valid science", because they are peer reviewed by people with qualifications, and so on.