Abstract
When random variables are used to represent variability, the risk equation has mathematical properties poorly understood by many risk assessors, variability represents the heterogeneity in a well‐characterized population, usually not reducible through further measurement or study. We follow the lead of most mathematicians in using random variables to represent and analyze variability. To illustrate the issues, we use LogNormal distributions to model variability.
Notes
P.O. Box 382669 Harvard Square Station, Cambridge, MA 02238–2569; Tel: (617) 864–4300; Fax: (617) 864–9954; e‐mail: [email protected]
245 Summer Street, Boston, MA 02210; Tel: (617) 589–1765; Fax: (617) 589–2922; e‐mail: [email protected]