Variance Structure
The variance functions are used to model heteroscedasticity in the within-group error.
Type The type of within-group variance function. The available functions are:
None, the default
Exponential, exponential of a covariate (varExp)
Power, power of a covariate (varPower)
Constant Plus Power, constant plus power of a covariate (varConstPower)
Identity, different variances per level of a factor (varIdent)
Fixed, fixed weights, determined by a covariate (varFixed)
Combination, a combination of variance functions (varComb)
See the online Help for the corresponding varFunc S-PLUS function to determine the specific parameters for each of the variance functions.
Covariate Choose the variance covariate from the dropdown list.
Group Variables Choose the grouping factor(s) from the dropdown list. Multiple grouping factors are converted to ordered factors.
Parameters Enter parameters as needed for the different variance functions.