You can do quite a bit without making strong distributional assumptions. There are nonparametric tests for all sorts of things (and the basic ones are covered i

You can do quite a bit without making strong distributional assumptions. There are nonparametric tests for all sorts of things (and the basic ones are covered in good first intro to stats books too).