Why (Experimentally) Is a Set of Quantons Prepared in a Superposition

Multiple Choice

Why (experimentally) is a set of quantons prepared in a superposition of a given observable's eigenvectors different from a classical mixture of quantons that have "already collapsed" to the different eigenvectors (with the same probabilities as predicted by the superposition) ?

A) We can (in appropriate circumstances) show that the superposition is an eigenvector of some other observable, but this is not possible for a mixture.
B) We can construct situations where quantons in the superposition state exhibit interference effects that cannot be observed for a mixture.
C) Both A and B.
D) Neither A nor B (explain) .