Columns in a data table can contain special column properties. Column Properties Menu in the Column Info Dialog shows that a column called Stretch has two special properties: Role and Response Limits, that were assigned by the Custom Designer when the table was created. To see the example in Column Properties Menu in the Column Info Dialog, open Bounce Data.jmp from the Design Experiment folder found in the sample data installed with JMP. Then, right-click the column name in the data table and select Column Info. When the Column Info dialog appears, click on the property you want to see.
Column Properties Menu in the Column Info Dialog
All special column properties are covered in the Using JMP. The following discussion gives details about properties specific to DOE and that are useful for analyzing DOE data.
For continuous variables, the Coding column property transforms data in the range you specify from –1 to +1. When you analyze the coded variable, JMP uses those transformed data values to compute meaningful parameter estimates. You can specify the range in which the low and high values of the column are transformed.
Coding Column Property in Column Info Dialog
To set up the CuS04 column as a mixture factor, first select Cols > Column Info to see the Column Info dialog for CuS04. Then continue as follows:
1.
Select Mixture from the Column Properties drop-down menu. Upper and Lower limits, and the sum of the limits appear in a panel on the dialog, as shown in Column Info to Create Mixture Column For Analysis. You can use these limits, or enter your own values.
2.
Optionally, check the boxes beside L PseudoComponent Coding, U PseudoComponent Coding, or both L and U PseudoComponent Coding. Using the example in Column Info to Create Mixture Column For Analysis, where the mixture sum value is 1, the terms are coded as:
for the L pseudocomponent
for the U pseudocomponent
where Li and Ui are the lower and upper bounds, L is the sum of Li and U is the sum of Ui.
Note: If you check either L PseudoComponent Coding or U PseudoComponent Coding for the mixture coding of one mixture factor and you check the other alternative for one or more other mixture factors in the model, of if you check both boxes for one or more of the mixture factors, the Fit Model platform uses the L coding if (1 – L) < (U – 1), and the U coding otherwise. If only one coding box is checked consistently for all mixture factors in the model, then only that one pseudocomponent coding is used.
3.
Select the Design Role Column Property, and choose Mixture from its drop down menu.
4.
Click OK. The properties icon () now appears next to the column name in the columns panel, indicating the column contains one or more column properties.
Column Info to Create Mixture Column For Analysis
Completed Response Limits shows the panel with values that specify lower, middle, and upper limits, and a desirability value. You can also select a possible goal for a DOE response variable: Maximize, Match Target, Minimize, or None. If you have more than one response, you can enter an importance value, which lets JMP know how to weigh the importance of one response against another.
1.
Click the column name Stretch in the data grid and select Cols > Column Info. The Column Info dialog appears.
2.
Select Response Limits from the Column Properties drop-down menu.
4.
When you have two responses, enter a number in the Importance box to indicate the amount of weight you want this response to have when JMP computes the overall desirability.
Completed Response Limits
6.
Click OK. The properties icon () now appears next to the column name in the column panel of the data table to indicate that the column contains a property.
1.
Click the column name in the data grid then select Cols > Column Info. The Column Info window appears.
2.
Select Design Role from the Column Properties drop-down menu, as shown in Assign a Design Role to a Factor Column. Design role information appears on the right.
3.
Click the Design Role drop-down menu and select how you want JMP to use the factor column: Continuous, Categorical, Blocking, Covariate, Mixture, Constant, Signal, Noise, Uncontrolled, Random Block, or Discrete Numeric.
4.
Click OK to see the property icon () next to the column name in the data table’s column panel.
Assign a Design Role to a Factor Column
To create split plot or split-split plot designs, you must identify a factor as having values that are hard to change, or very hard to change. This is done in the DOE design panel (see Creating Split Plot Designs, for details) each time you design an experiment. However, if you know that every time you use that factor, you want it to be considered hard or very hard to change, you can save yourself steps by setting up a column property to be used in all experiments using that factor. To do this:
1.
Click the column name in the data grid and then select Cols > Column Info to see the Column Info dialog for that column.
2.
Select Factor Changes from the Column Properties drop-down menu, as shown in Factor Changes Column Property.
3.
Click the Factor Changes button and select Easy, Hard, or Very Hard from the Factor Changes drop-down menu.
4.
Click OK. The properties icon () now appears next to the column name in the column panel of the data table.
Factor Changes Column Property