S. HERNBERG. (1981). Evaluation of “negative” results in occupational epidemiology