The discussion paper provides estimates of mode-specific measurement biases for key health statistics between web and interviewer-assisted data collection. Biases for some statistics are sizeable and ...