Bill's team is concerned that there may be bias in which data points are chosen if the team hand-selects them. To combat this concern, Bill believes that random sampling may be a good approach. Is this a sound idea?