Low impact journals may be significantly worse, but if fewer people look for defects fewer defects get found.