Do Name-Based Treatments Violate Information Equivalence? Evidence from a Correspondence Audit Experiment