 |
 |
Dear list: I am having the same problem as Jianbin Fu. In a model with fixed variances of latent variables (LVs), the correlations among the LV's are to be estimated. The variances of the LVs are fixed to ensure scale identification. The usual alternative, fixing the relationship between each LV and one of the response variable, seems to work well in linear factor analysis. However, with some non-linear response functions, fixing measurement relations no longer appears to be straightforward. Would anyone be able to share advice about useful priors for correlation (not covariance) matrices and how to implement these in WinBUGS, assuming this is at all possible? Thank you, Werner Wothke Research, Quantitative Methods CTB/McGraw-Hill 20 Ryan Ranch Road Monterey, CA 93940 831-393-6472 -----Original Message----- From: Jianbin Fu [mailto:[log in to unmask]] Sent: Tuesday, June 24, 2003 9:07 AM To: [log in to unmask] Subject: [BUGS] correlation matrix in multivariate normal dist. Hi folks, I am trying to use the correlation matrix, instead of covariance matrix, in the multivariate normal dist. (MVN). I have tried to first sample the prior covariance matrix from a Wishart, and then convert it to the correlation matrix. But it did not work because in Winbugs 1.4 Wishart can only be used as the prior distribution of covariance matrix in MVN, otherwise Winbugs can not find a sampler for Wishart. The reason that I want to use correlation matrix is to avoid the scale indeterminacy in latent trait models. Any suggestion is highly appreciated on how to model the prior distribution of correlation matrix in MVN in Winbugs as well as how to solve the scale indeterminacy in latent trait models when using MCMC. Jianbin ------------------------------------------------------------------- This list is for discussion of modelling issues and the BUGS software. For help with crashes and error messages, first mail [log in to unmask] To mail the BUGS list, mail to [log in to unmask] Before mailing, please check the archive at www.jiscmail.ac.uk/lists/bugs.html Please do not mail attachments to the list. To leave the BUGS list, send LEAVE BUGS to [log in to unmask] If this fails, mail [log in to unmask], NOT the whole list ------------------------------------------------------------------- This list is for discussion of modelling issues and the BUGS software. For help with crashes and error messages, first mail [log in to unmask] To mail the BUGS list, mail to [log in to unmask] Before mailing, please check the archive at www.jiscmail.ac.uk/lists/bugs.html Please do not mail attachments to the list. To leave the BUGS list, send LEAVE BUGS to [log in to unmask] If this fails, mail [log in to unmask], NOT the whole list