Crossed categorization outside the lab: Findings from a factorial survey experiment