I don't think the claim relies on siblings' environments being the exact same. The point is just that the degree of intra-family similarity is almost always far greater than the degree of average inter-family similarity. To a reasonable approximation it's fair to call it 'the same' in the context of the comparison with non-siblings.