Many contemporary physicists suggest that we should take the claim that we live in a multiverse seriously. When doing this they often invoke arguments based on the controversial anthropic principle. Critics argue that this leads to untestable and unscientific theories. In this paper criteria are suggested that need to be satisfied before a multiverse theory should be considered scientifically respectable. One important point is that the multiverse is described in sufficient detail. In some cases, a multiverse theory can be testable; however, to properly test a multiverse theory it is important to distinguish new predictions from explanations based on the multiverse.