Learn R Programming

RcmdrPlugin.DoE (version 0.12-5)

Menu.FacDetailsGenTab: How to use the factor details tab for general designs

Description

This help file describes how to input factor details for general designs

Arguments

Inputs on tab Factor Details for all such menus

Modify factor details for selected factor

is the frame in which changes can be made. A factor can be selected by clicking on the arrow below Select (and also by clicking on an item in the listboxes below the frame). Repeatedly jumping through the frame with the Tab key automatically cycles through the factors one by one.

The number of levels MUST BE ENTERED for each factor. (Jumping out of the entry box without entering a valid number of levels creates an error message popup.) The levels themselves are then automatically set to the integer numbers from one to the number of levels. Different levels (separated by blanks) can be manually entered. It is recommended to use the automatic levels for preliminary checks on available designs and their properties, but to enter meaningful factor levels for experiments that will actually be run.

The entry Comment or label is included into an exported html file as annotation for the factor. This may be useful for conducting the experiment.

Buttons for changing the factor order

If factor details have been entered and it is discovered later that the factors are needed in a different order (e.g. because of the alias structure of a regular 2-level fractional factorial design), the buttons Move Up and Move Down can be used for changing the order of the factors.

Manually specify column numbers for array

is a checkbox for the orthogonal array menu that is available if a specific design has been selected by ID (from a drop-down menu on the Base Settings tab). If this box is checked, the factor details include an additional column for the column numbers that needs to be filled with a valid column number for each factor. The number of levels is then automatically derived.

Specifying column numbers can be useful, if an expert user has investigated the properties of a design and wants to obtain a better-suited design than the default one. For example, when accomodating two 2-level factors, one 4-level factor and one 8-level factor in L64.2.5.4.10.8.4, it is better to use columns 2 and 3 than the default columns 1 and 2 for the two 2-level factors. With the latest version of the software, this optimization can also be done automatically using the setting “min3” for Automatic Optimization on the “Base Settings” tab; nevertheless, because of performance reasons, it may be wise to specify known good column numbers, if available. Note that explicit specification of column numbers overrides any optimization request.

Author

Ulrike Groemping