Analysis of Variance by ANOVA, Regression or REML Options

Use this to select the output to be generated initially by an analysis of variance by ANOVA, Regression or REML.

Display

Specifies which items of output are to be produced.

AOV table	Analysis of variance table
Means	Tables of means for treatment terms
Residuals	Tables of estimated residuals
Information	Details of large residuals, non-orthogonality and any aliasing in the model
F-probability	F-probabilities for the analysis of variance table

Standard errors

When printing tables of means, you can specify the form in which standard errors will be displayed: as standard errors of the means, standard errors of differences of means (SEDs), or least-significant differences (LSDs). If required, you can print more than one type of standard error. The significance level for LSDs can be specified as a percentage (default 5) in the accompanying field. Note that LSDs cannot be formed for tables combining information over several strata.

Available data

This lists data structures appropriate to the current input field. The contents will change as you move from one field to the next. Double-click a name to copy it to the current input field or type the name.

Weights

A variate of weights can be supplied to give varying influence of each unit on the fit of the model. This would usually correspond to a known pattern of variance of the observations, where the weights would be the reciprocals of the variances. A variate that specifies the weights in the analysis can be selected by double clicking an item in Available data list or you can type its name into the field. If the y-variate has been specified, only variates that match its length will be displayed in the list.

Factorial limit on treatment terms

Controls the level of interactions to be fitted. The value entered indicates the level of interaction allowed in the model (that is, set a limit on the maximum number of factors in the interaction terms that are fitted). For example, setting this to 1 will fit only the main effects (1 factor only in fitted terms), and setting this to 2 will only allow first order interactions of the form A.B, and higher order interactions such as A.B.C will not be included in the model. The terms not included in the model will be assigned to the error term.

Limit on loss of efficiency for regression

Use this to limit the maximum loss of efficiency occurring on any treatment contrast, if the analysis is done by regression. If more efficiency than this amount is lost, the analysis will switch from a regression analysis to a REML analysis.