Theoretically, random assignment should eliminate

A. the need to use statistics.

B. concerns over validity.

C. many confounding variables.

D. the need for a representative sample