When data that are frequencies (counts) are listed in several columns of your data table, you must transform the data into the form that you need for logistic regression. For example, the Ingots2.jmp data table (see Figure 10.11) has columns Nready and Nnotready. These columns give the number of ingots that are ready to roll and ingots that are not ready to roll for each combination of Heat and Soak values.
Figure 10.11 Ingots2.jmp Sample Data Table
Before fitting a logistic regression model, use the following steps to stack the Nready and Nnotready columns into a single column:
1.
|
2.
|
Select Tables > Stack.
|
3.
|
4.
|
Click OK.
|
This creates the new table in Figure 10.12. Label is the response (Y) column and Data is the frequency column.
This stacked data table is equivalent to the Ingots.jmp sample data table used in Example of Nominal Logistic Regression.
Figure 10.12 Stacked Data Table