Email discussion lists for the UK Education and Research communities

## BUGS@JISCMAIL.AC.UK

#### View:

 Message: [ First | Previous | Next | Last ] By Topic: [ First | Previous | Next | Last ] By Author: [ First | Previous | Next | Last ] Font: Proportional Font

#### Options

Subject:

Multivariate normal of data and parameters

From:

Date:

Fri, 26 Oct 2007 18:04:23 +0100

Content-Type:

text/plain

Parts/Attachments:

 text/plain (43 lines)
 ```Dear All, I am using winbugs to fit a joint model of a set of weight measurements over time with two adult outcomes BMI and plasma glucose for each individual. Interest lies in quantifying the correlation of the adult outcomes with the growth parameters. The growth model is parameterized as a simple 2-level (individual-occassion) polynomial function of the occasion timing and the intercepts and slopes are random parameters. I am not quite sure what's the best way to proceed with the joint modelling of the two adult outcomes (2-level or individual outcomes) with the growth parameters (2-level individual level parameters) I have tried 2 different ways: 1. specify a multivariate normal (MVN) for the outcomes and the parameters with a wishart-prior for the variance-covariance matrix as specified in mlwin Winbugs proceeds with the estimation, although that DIC is greyed out. So, I am a bit worried about what's going on. 2. Specify univariate normals for the adult outcomes and then specify an MVN for the means of these two outcomes and the growth parameters. Now, winbugs is happier and there DIC is calculated - I do have though the added complication to specify a prior for the precision of the means of the two adult outcomes. In other words, now I am specifying a measurement error model for the adult outcomes. So, for the prior for these precision parameteres I could either give a fixed value (something large) or a gamma(0.001,0.001). The second of-course won't work very well, slow convergence because there is no information in the data for this. For the first specification however, I was expecting not to be very different from 1. But there are quite large differences. There are 1000 individuals with 8459 observations. I was wondering if anybody has any thoughts on how's best to proceed. Many thanks Daphne ------------------------------------------------------------------- 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 ```