"Triple blinding in epidemiologic studies"
Sir: I am aware of a third type of blinding (in addition to the joke that defines triple blinding as that when neither the patient, nor the observer, nor the researcher know what's going on). Some researchers go as far as requiring that the statistician is also kept blind for data analysis. Considering that proper data quality control and data management have been assured, and that statistical analysis is guided by research questions and hypotheses, this practice seems rather absurd and should be avoided.
Hector E. Velasco-Mondragon, M.D., M.Sc.
Candidate in Epidemiology
Johns Hopkins School of Hygiene and Public Health
Centro de Investigaciones en Salud Poblacional
Instituto Nacional de Salud Publica
Cuernavaca, Mor. Mexico
Competing interests: No competing interests