Analysis of Variance

The General Analysis of Variance menu allows you to produce analyses of variance for a wide variety of designs, ranging from simple one-way ANOVA to Latin squares. The General Analysis of Variance option can be used to fit any generally balanced design not included in the more specific selections. Unbalanced data sets can be analysed using the unbalanced treatment design, or alternatively by using regression (single error term) or REML (more than one error term). Each of the designs, with the exception of unbalanced treatment structure, can also be generated using the Generate a Standard Design menu.

Design

Selects the type of design to be analysed. See the help specific to each design for a description of its features and the information to be entered.

Available Data

This lists data structures appropriate to the current input field. The contents of this box will change as you move from one field to the next. You can double-click a name in the list box to copy it to the current input field.

Y-Variate

This should be set to the variate containing the data values. Available Data lists variates available for analysis.

Treatment Structure

The treatment structure is specified according to the type of design selected. See the help for the chosen type of design for further details.

Contrasts

The contrasts button will open a dialog allowing you to set up contrasts providing comparisons of the effects produced at different levels of a treatment factor.

Block Structure

The block structure is specified according to the type of design selected. See the help for the chosen type of design for further details.

Interactions

For designs with more than one treatment factor you can specify whether interactions can be fitted in addition to main effects.

Covariate

Select the Covariate check box on the Analysis of Variance menu if you want to perform analysis of covariance. This produces an additional box in which you can supply a list of the covariates for the analysis. If there is more than one, their names can be separated by spaces or commas. Alternatively, you can specify a model formula to define the terms to be fitted in the covariate analysis.

Options

This controls display options, allowing you to select the output to be generated initially by ANOVA - the same information can also printed after the analysis, using Further Output.

Save

This allows you to specify data structures in which to save fitted values, residuals and treatment means from the analysis.

Further Output

This allows you to display additional results from the analysis of variance and produce graphical output of fitted treatment means and diagnostic plots.