To try to make experimental groups as equivalent as possible before beginning any manipulations, one should use:

Group of answer choices

random selection

internal validity

random assignment

placebo group