On the second page, they say they chose participants who knew each other to "prevent unfamiliarity from affecting the form of collaboration" but that seems odd to me. Wouldn't familiarity be just as likely, or more likely to affect the form of collaboration?