-------- Original Message -------- Subject: Find outliers in MorphoJ Date: Thu, 1 Mar 2012 14:40:18 -0500 From: Steve Wang <[email protected]> Reply-To: [email protected] To: [email protected] Hello, This may be an amateurish question with regard to the "Find Outliers" option in MorphoJ. I read the online user guide for this specific option and know that a graph with Procrustes distance (absolute shape deviation) is generated when you have a small sample, and a Mahalanobis distance graph (relative shape deviation) is generated for a larger dataset--kind of an automatic thing. Now, will the two graphs look very different, if one can choose? I am very curious because most of the stats sites suggest using Mahalanobis distances to look for multivariate outliers. Steve -- Steven L. Wang Visiting Researcher | Lecturer School of Anthropology University of Arizona Telephone: (520) 621-2540 E-mail:**[email protected] <mailto:[email protected]>
