Do group responses mask the effects of air pollutants on potentially sensitive individuals in controlled human exposure studies?