Each person in a simple random sample of 1,500 received a survey, and 412 people returned their survey. How could nonresponse cause the results of the survey to be biased?