Variance Components

Creates variance components for ANOVA/ANCOVA designs with random effects using various estimation techniques, including ANOVA (and denominator synthesis), Minimum Variance Quadratic Unbiased Estimators method (MIVQUE(0)), Restricted Maximum Likelihood (REML), and Maximum Likelihood (ML) estimation. Note the assignment of variables: Continuous dependent - Dependent Categorical dependent - Random effects Categorical predictors - Fixed effects Continuous predictors - Covariates The General Linear Models (GLM) procedures will also analyze and compute the ANOVA results with synthesized error terms. The aforementioned analysis more specifically involves ANOVA/ANCOVA designs of arbitrary complexity with random effects and estimate variance components (using the ANOVA method)

General

Element Name Description
Detail of computed results reported Specifies the detail of computed results reported. If Minimal detail is requested, only the final estimates for the variance components are reported; if All results is requested, various additional matrices and graphs will be reported.
Model Specifies the model to use for fitting responses on the dependent variable(s) and to specify the order, or degree, of interactions of the factors to include in the model. Select the Factorial design option to perform the analysis using a factorial design based on combinations of the levels of the factors in the analysis. Select the Hierarchically nested design option to perform the analysis using hierarchically nested effects. The order in which factors are selected determines the order of nesting of effects. For example, if three factors A, B, and C are selected, in that order, the model for the analysis would include effects for C nested within B, B nested within A, and the A main effect.
Factorial to degree Only applicable if Factorial design is specified; Statistica will perform the analysis using a factorial design based on combinations of the levels of the factors in the analysis.
Fixed-random interactions are Specifies whether the interactions between the fixed factors and the random factors are to be treated as fixed effects or random effects; this option is only applicable if the Factorial design model is selected. Note that interactions between fixed factors are always treated as fixed effects, and that interactions between random factors are always treated as random effects.
Interpretation of codes Specifies how to interpret the codes for the nested effects; this option is only applicable if the Hierarchically nested design model is selected. When the Consecutive overall levels option is selected, Statistica expects unique codes for each level of the nested effects. When the Levels within other factors option is selected, the program expects the codes to reflect the levels nested within the other factors.

Variance Components

Element Name Description
Method Choose the method to be used for estimating the variance components for the random effects; available options are Expected Mean Squares (ANOVA estimates), MIVQUE 0, Restricted Maximum Likelihood (REML), and Maximum Likelihood (ML).
SS type, for ANOVA Choose the type of sums of squares decomposition when estimating the variance components for the random effects in the model using Expected MS (mean squares) as the Method (see above). This option will also affect how the overall ANOVA table will be computed for unbalanced or incomplete designs.