Estimation of constant-CV regression models
A typical formulation for a linear mixed model is Y = X(be) + Z(u) where (be) is a vector of "fixed" parameters, (u) is a vector of "random effects", and X and Z are matrices whose columns consist of design variables and/or covariates. In many applications the variances of the random effects corresponding to a particular column of Z are not constant, but are more realistically modeled as being proportional to some power of E(Y) the mean of Y. In particular, if the variance is proportional to the square of E(Y), we have a constant CV model. This talk will give an example of such a model and show how -xtmixed- can be used to estimate it and do proper inference on the estimated parameters. Results will be compared with Bayesian estimation under WINBUGS.
|Date of creation:||29 Jul 2008|
|Date of revision:||28 Aug 2008|
|Contact details of provider:|| Web page: http://stata.com/meeting/snasug08/|
More information through EDIRC
When requesting a correction, please mention this item's handle: RePEc:boc:nsug08:4. See general information about how to correct material in RePEc.
For technical questions regarding this item, or to correct its authors, title, abstract, bibliographic or download information, contact: (Christopher F Baum)
If references are entirely missing, you can add them using this form.