Assessing observer variability: Authors' reply