Data that are frequencies (counts) listed in several columns of your data table are not the form that you need for logistic regression. For example, the data table in the data folder (see Original Data Table) has columns Nready and Nnotready that give the number of ready and number of not ready ingots for each combination of Heat and Soak values. To do a logistic regression, you need the data organized like the table in Stacked Data Table.
To make a new table, suitable for logistic regression, select the Stack command from the Tables menu. Complete the Stack dialog by choosing Nready and NNotReady as the columns to stack, and then click OK in the Stack dialog. This creates the new table in Stacked Data Table. If you use the default column names, Label is the response (Y) column and Data is the frequency column.
The example in the section Introduction to Logistic Models, shows a logistic regression using a sample data table It has a frequency column called count (equivalent to the Data column in the table below) and a response variable called Ready, with values 1 to represent ingots that are ready and 0 for not ready.
Original Data Table
Stacked Data Table