Information Equivalence in Survey Experiments