-------- Original Message -------- Subject: Repeatability by PC Score Date: Wed, 28 Jan 2009 17:04:01 +1100 From: Paul Sanfilippo <[email protected]> To: [email protected] Hi all, I hope this isn't a silly question, but I'd appreciate some feedback none the less. I'm wanting to calculate intra- (one observer, one specimen image, 2 measurements) and inter- (two observers, one specimen image, 1 measurement each) observer variability using a sample of my dataset and I'm planning on running the data through several techniques (EFA, SSL, eigenshape analysis) for preliminary comparisons as part of this. MorphoJ offers a nice way of doing this (if I understand correctly) by performing a Procrustes ANOVA, but I'm not dealing purely with landmarks, hence the examination of other methods. I came across a paper: Kieser et al. The uniqueness of the human anterior dentition a geometric morphometric analysis. J Forensic Sci. 2007 where the investigators used the score on PC 1 for each specimen in order to calculate an intraclass correlation coefficient. In order to keep things consistent in terms of data analysis across the various techniques, this seems quite appealing. What are people's thoughts on doing this - is it an acceptable (albeit quick and easy) way of assessing repeatability. Thank you. Paul Sanfilippo Centre for Eye Research Australia University of Melbourne -- Replies will be sent to the list. For more information visit http://www.morphometrics.org
