vince, you are correct. and so is the affy code. thanks, rafael ---------- Forwarded message ---------- Date: Tue, 06 May 2003 07:52:14 -0400 (EDT) From: Vincent Carey 525-2265 <stvjc@channing.harvard.edu> To: rafa@jhu.edu Subject: scaling the mad hi -- your e-mail stated that 1.48*mad is unbiased for variance ... but it is actually unbiased for the std deviation. just a typo... the rest of the text is accurate.