1 “... on the observational effects that these models generate, accounting for tests and comparisons which can be made on all...”