Analyze Factorial Design - Storage
main topic     see also      

Stat > DOE > Factorial > Analyze Factorial Design > Storage

You can store the residuals, fitted values, and many other diagnostics for further analysis (see Checking your model). Minitab stores the checked values in the next available columns and names the columns

Dialog box items

Fits: Check to store the fitted values. One column is stored for each response variable.

Residuals: Check to store the residuals. One column is stored for each response variable.

Standardized residuals: Check to store the standardized residuals.

Deleted residuals: Check to store Studentized residuals.

Whole plot residuals: Check to store whole plot residuals for 2-level split-plot designs.

Leverages: Check to store the leverages. This dialog box item is not available for split-plot designs.

Cook's distance: Check to store Cook's distance. This dialog box item is not available for split-plot designs.

DFITS: Check to store the DFITS.

Effects: Check to store the effects. Minitab stores one column for each response. These are the effects that are printed on the output.

Coefficients: Check to store the estimated coefficients of the regression equation. These are the same coefficients as are printed in the output. If some terms are removed because the data cannot support them, the removed terms do not appear on the output.

Design matrix: Check to store the design matrix corresponding to your model. If the design was blocked into k blocks, there are (k-1) columns for block dummy variables. Analyze Factorial Design uses the same method of coding blocks as General Linear Model. The block dummy variables are followed by one column for each component. When the model has 2-way interactions, the design matrix contains a column for each interaction. The column for a 2-way interaction is the product of the corresponding two components. When the model has 3-way interactions, the design matrix contains a column for each interaction. The column for a 3-way interaction is the product of the corresponding three components.

Means: Check to store the least squares means that you specified in the Options subdialog.

Box-Cox Transformation: If you transform your response variable, you can store information about the transformation.

Box-Cox transformation of response: Check to store the transformed values of the response variable.

Fits for original response: Check to store the fits for the original response.

Means for original response: Check to store the least squares means that you specified in the Options subdialog.