Specify Multiple Subsets

Graphical Analytic Techniques

This dialog is accessible by clicking either the Mark Selected Subsets or Specify Subsets buttons. In either case, this dialog is used to specify the selection conditions which will subset or categorize (respectively) the cases in a data file to be assigned to different graphs (in categorized graphs), categories (e.g., histograms or box plots), or marked classes (e.g., in multiple subset scatterplots). The subsetting conditions that you specify here will be used as legends of the respective plots. See also the Specifying Multiple Subsets example.

Non-exclusive categories. Even though you can specify non-exclusive categories, when the selection (categorization) conditions are processed, they will always assign each case only once (to the first category where it "fits"). Thus, multiple subsets or categories created via this option will always be mutually exclusive.

Subset. Use up to 16 categories to subset or categorize a variable. However, remember that each case will only be assigned to a category once (resulting in mutually exclusive categories); thus, if a case fits a particular selection condition (e.g., Subset 3), but it also fits a previous condition (e.g., Subset 2), then the case will not be assigned to Subset 3, but only to Subset 2.

Include, Exclude if. When you choose Include if and the subset condition evaluates to true, the values of the selected variable for cases (observations) that meet the condition will be placed in the category (or subset) defined by that condition. Exclude if will exclude the value from the category if the case meets the condition.

Include, Exclude cases. When you choose Include cases and the subset condition evaluates to true, the values of the specified cases (observations) will be placed in the category (or subset) defined by that condition. Exclude cases will exclude the value from the category for the specified cases.

Subset condition. Specify the subset condition in the blank Subset line following the standard STATISTICA Case Selection Conditions.

· Operators: =, <>, <, >, <=, >=, NOT, AND, OR

· Variable Names: Specify either variable numbers (e.g., v1, v2, v3, . . . ) or variable names (e.g., GENDER, Date, time, . . .).

· Case Number: v0

· Examples: v1=0 OR v2>=0;

(v1<1 OR v9='YES' ) AND v4<>0

Notes. Be sure to use single quotation marks to reference the text values of variables, e.g., 'Yes'. In order to reference date values, use the Julian value representing that date (e.g., use the Julian date 24858 instead of 1/21/1968; see How are dates represented in STATISTICA data files?).

Since the text specified here will be used as plot legends, you can use the shortcut symbols for certain operators in order to save space when specifying the selection conditions. Use the following shortcut symbols:

& for the operation AND

! for the operation OR

# for the operation NOT EQUAL

~ for the operation NOT

Open. This button opens the Open Conditions dialog from which you can select a file (containing all of the subsetting categories) to open into the Specify Multiple Subsets dialog. Files opened from this dialog are standard text files with the file name extension *.txt.

Save As. This button opens the Save Conditions dialog from which you can designate a file into which all of the subsetting conditions can be saved. This file will be saved in text file format (*.txt extension).