In the Context of Experimental Research, the Logic of Random

In the context of experimental research, the logic of random assignment is

A) if participants in an experiment are assigned to each group by stratification, the actual differences between the groups will sustain over the long run.
B) if participants in an experiment are assigned to each group only by chance, the actual similarities between the groups will sustain over the long run.
C) if participants in an experiment are assigned to each group by stratification, the potential differences between the groups on other characteristics will cancel out over the long run.
D) if participants in an experiment are assigned to each group only by chance, the potential differences between the groups on other characteristics will cancel out over the long run.