Reporting health events in household interviews: Effects of reinforcement, question length, and reinterviews