Dear all, i'm interested in references regarding specifying (unbalanced) multivariate mixed regression models in univariate format with more than one random effect. R. Thompson (1973) biometrics 29:527-550 and K.Meyer Biometrics 41:153-165 have done work on the above area with only one random effect, they derive the mixed model equations and parameter estimates for this specification. But i've not found any work on more than one random effect and heteroscedasticity with this univariate specification. Also, would like information on multivariate F-tests with kenwardroger's denominator degrees of freedom. Khuri (1998) ,Statistical tests for mixed linear models, has done work on F-tests with satterwaite's ddfm but not KR's ddrm.
Thanks for any help. Hars . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
