Using random assignment to groups in an experiment promotes all of the following EXCEPT:

equivalent groups

internal validity

reduced sampling bias

reduced selection bias