Random assignment to conditions is the defining feature of experimental design, which allows causal inference.