Visual binning

6/17/13 Visual Binning
publib.boulder.ibm.com/infocenter/spssstat/v20r0m0/advanced/print.jsp?topic=/com.ibm.spss.statistics.help/idh_bander_gating.htm&pageBreak=true&breakConfi… 1/9
Visual Binning
Contents
1. To Bin Variables
2. Binning Variables
3. Automatically Generating Binned Categories
4. Copying Binned Categories
5. User-Missing Values in Visual Binning

Visual Binning
Visual Binning is designed to assist you in the process of creating new variables based on
grouping contiguous values of existing variables into a limited number of distinct categories. You
can use Visual Binning to:
• Create categorical variables from continuous scale variables. For example, you could use a
scale income variable to create a new categorical variable that contains income ranges.
• Collapse a large number of ordinal categories into a smaller set of categories. For example, you
could collapse a rating scale of nine down to three categories representing low, medium, and
high.
In the first step, you:
Select the numeric scale and/or ordinal variables for which you want to create new categorical
(binned) variables.
Optionally, you can limit the number of cases to scan. For data files with a large number of
cases, limiting the number of cases scanned can save time, but you should avoid this if possible
because it will affect the distribution of values used in subsequent calculations in Visual Binning.
Note: String variables and nominal numeric variables are not displayed in the source variable list.
Visual Binning requires numeric variables, measured on either a scale or ordinal level, since it
assumes that the data values represent some logical order that can be used to group values in a
meaningful fashion. You can change the defined measurement level of a variable in Variable View
in the Data Editor. See the topic Variable measurement level for more information.
© Copyright IBM Corporation 1989, 2011.

1. To Bin Variables
From the menus in the Data Editor window choose:
Transform > Visual Binning...
(binned) variables.
Note: String variables and nominal numeric variables are not displayed in the source variable list.
Visual Binning requires numeric variables, measured on either a scale or ordinal level, since it
assume that the data values represent some logical order that can be used to group values in a
meaningful fashion.
Select a variable in the Scanned Variable List.
Enter a name for the new binned variable. Variable names must be unique and must follow variable
naming rules. See the topic Variable names for more information.
Define the binning criteria for the new variable. See the topic Binning Variables for more
information.
Click OK.

2. Binning Variables
The Visual Binning main dialog box provides the following information for the scanned variables:
Scanned Variable List. Displays the variables you selected in the initial dialog box. You can sort
the list by measurement level (scale or ordinal) or by variable label or name by clicking on the
column headings.
Cases Scanned. Indicates the number of cases scanned. All scanned cases without user-
missing or system-missing values for the selected variable are used to generate the distribution
of values used in calculations in Visual Binning, including the histogram displayed in the main
dialog box and cutpoints based on percentiles or standard deviation units.
Missing Values. Indicates the number of scanned cases with user-missing or system-missing
values. Missing values are not included in any of the binned categories. See the topic User-
Missing Values in Visual Binning for more information.
Current Variable. The name and variable label (if any) for the currently selected variable that
will be used as the basis for the new, binned variable.
Binned Variable. Name and optional variable label for the new, binned variable.
• Name. You must enter a name for the new variable. Variable names must be unique and must
follow variable naming rules. See the topic Variable names for more information.
• Label. You can enter a descriptive variable label up to 255 characters long. The default
variable label is the variable label (if any) or variable name of the source variable with (Binned)
appended to the end of the label.
Minimum and Maximum. Minimum and maximum values for the currently selected variable,
based on the scanned cases and not including values defined as user-missing.
Nonmissing Values. The histogram displays the distribution of nonmissing values for the
currently selected variable, based on the scanned cases.
• After you define bins for the new variable, vertical lines on the histogram are displayed to
indicate the cutpoints that define bins.
• You can click and drag the cutpoint lines to different locations on the histogram, changing the
bin ranges.
• You can remove bins by dragging cutpoint lines off the histogram.
Note: The histogram (displaying nonmissing values), the minimum, and the maximum are based on
the scanned values. If you do not include all cases in the scan, the true distribution may not be
accurately reflected, particularly if the data file has been sorted by the selected variable. If you
scan zero cases, no information about the distribution of values is available.
Grid. Displays the values that define the upper endpoints of each bin and optional value labels
for each bin.
• Value. The values that define the upper endpoints of each bin. You can enter values or use
Make Cutpoints to automatically create bins based on selected criteria. By default, a cutpoint
with a value of HIGH is automatically included. This bin will contain any nonmissing values
above the other cutpoints. The bin defined by the lowest cutpoint will include all nonmissing
values lower than or equal to that value (or simply lower than that value, depending on how
you define upper endpoints).
• Label. Optional, descriptive labels for the values of the new, binned variable. Since the values

of the new variable will simply be sequential integers from 1 to n, labels that describe what
the values represent can be very useful. You can enter labels or use Make Labels to
automatically create value labels.
To Delete a Bin from the Grid
Right-click on the either the Value or Label cell for the bin.
From the pop-up context menu, select Delete Row.
Note: If you delete the HIGH bin, any cases with values higher than the last specified cutpoint
value will be assigned the system-missing value for the new variable.
To Delete All Labels or Delete All Defined Bins
Right-click anywhere in the grid.
From the pop-up context menu select either Delete All Labels or Delete All Cutpoints.
Upper Endpoints. Controls treatment of upper endpoint values entered in the Value column of
the grid.
• Included (<=). Cases with the value specified in the Value cell are included in the binned
category. For example, if you specify values of 25, 50, and 75, cases with a value of exactly
25 will go in the first bin, since this will include all cases with values less than or equal to 25.
• Excluded (<). Cases with the value specified in the Value cell are not included in the binned
category. Instead, they are included in the next bin. For example, if you specify values of 25,
50, and 75, cases with a value of exactly 25 will go in the second bin rather than the first,
since the first bin will contain only cases with values less than 25.
Make Cutpoints. Generates binned categories automatically for equal width intervals, intervals
with the same number of cases, or intervals based on standard deviations. This is not available if
you scanned zero cases. See the topic Automatically Generating Binned Categories for more
information.
Make Labels. Generates descriptive labels for the sequential integer values of the new, binned
variable, based on the values in the grid and the specified treatment of upper endpoints
(included or excluded).
Reverse scale. By default, values of the new, binned variable are ascending sequential integers
from 1 to n. Reversing the scale makes the values descending sequential integers from n to 1.
Copy Bins. You can copy the binning specifications from another variable to the currently
selected variable or from the selected variable to multiple other variables. See the topic Copying
Binned Categories for more information.

3. Automatically Generating Binned Categories
The Make Cutpoints dialog box allows you to auto-generate binned categories based on selected
criteria.
To Use the Make Cutpoints Dialog Box
(binned) variables.
Click Continue.
Select (click) a variable in the Scanned Variable List.
Click Make Cutpoints.
Select the criteria for generating cutpoints that will define the binned categories.
Click Apply.
Note: The Make Cutpoints dialog box is not available if you scanned zero cases.
Equal Width Intervals. Generates binned categories of equal width (for example, 1–10, 11–20,
and 21–30) based on any two of the following three criteria:
• First Cutpoint Location. The value that defines the upper end of the lowest binned category
(for example, a value of 10 indicates a range that includes all values up to 10).
• Number of Cutpoints. The number of binned categories is the number of cutpoints plus one.
For example, 9 cutpoints generate 10 binned categories.
• Width. The width of each interval. For example, a value of 10 would bin age in years into 10-
year intervals.
Equal Percentiles Based on Scanned Cases. Generates binned categories with an equal
number of cases in each bin (using the aempirical algorithm for percentiles), based on either
of the following criteria:
• Number of Cutpoints. The number of binned categories is the number of cutpoints plus one.
For example, three cutpoints generate four percentile bins (quartiles), each containing 25% of
the cases.
• Width (%). Width of each interval, expressed as a percentage of the total number of cases.
For example, a value of 33.3 would produce three binned categories (two cutpoints), each
containing 33.3% of the cases.
If the source variable contains a relatively small number of distinct values or a large number of
cases with the same value, you may get fewer bins than requested. If there are multiple
identical values at a cutpoint, they will all go into the same interval; so the actual percentages
may not always be exactly equal.
Cutpoints at Mean and Selected Standard Deviations Based on Scanned Cases. Generates
binned categories based on the values of the mean and standard deviation of the distribution of
the variable.

• If you don't select any of the standard deviation intervals, two binned categories will be
created, with the mean as the cutpoint dividing the bins.
• You can select any combination of standard deviation intervals based on one, two, and/or
three standard deviations. For example, selecting all three would result in eight binned
categories--six bins in one standard deviation intervals and two bins for cases more than
three standard deviations above and below the mean.
In a normal distribution, 68% of the cases fall within one standard deviation of the mean; 95%,
within two standard deviations; and 99%, within three standard deviations. Creating binned
categories based on standard deviations may result in some defined bins outside of the actual
data range and even outside of the range of possible data values (for example, a negative salary
range).
Note: Calculations of percentiles and standard deviations are based on the scanned cases. If
you limit the number of cases scanned, the resulting bins may not contain the proportion of
cases that you wanted in those bins, particularly if the data file is sorted by the source variable.
For example, if you limit the scan to the first 100 cases of a data file with 1000 cases and the
data file is sorted in ascending order of age of respondent, instead of four percentile age bins
each containing 25% of the cases, you may find that the first three bins each contain only
about 3.3% of the cases, and the last bin contains 90% of the cases.

4. Copying Binned Categories
When creating binned categories for one or more variables, you can copy the binning
specifications from another variable to the currently selected variable or from the selected
variable to multiple other variables.
To Copy Binning Specifications
(binned) variables.
Click Continue.
Define binned categories for at least one variable--but do not click OK or Paste.
Select (click) a variable in the Scanned Variable List for which you have defined binned
categories.
Click To Other Variables.
Select the variables for which you want to create new variables with the same binned categories.
Click Copy.
or
Select (click) a variable in the Scanned Variable List to which you want to copy defined binned
categories.
Click From Another Variable.
Select the variable with the defined binned categories that you want to copy.
Click Copy.
If you have specified value labels for the variable from which you are copying the binning
specifications, those are also copied.
Note: Once you click OK in the Visual Binning main dialog box to create new binned variables (or
close the dialog box in any other way), you cannot use Visual Binning to copy those binned
categories to other variables.

5. User-Missing Values in Visual Binning
Values defined as user-missing (values identified as codes for missing data) for the source
variable are not included in the binned categories for the new variable. User-missing values for
the source variables are copied as user-missing values for the new variable, and any defined
value labels for missing value codes are also copied.
If a missing value code conflicts with one of the binned category values for the new variable, the
missing value code for the new variable is recoded to a nonconflicting value by adding 100 to the
highest binned category value. For example, if a value of 1 is defined as user-missing for the
source variable and the new variable will have six binned categories, any cases with a value of 1
for the source variable will have a value of 106 for the new variable, and 106 will be defined as
user-missing. If the user-missing value for the source variable had a defined value label, that
label will be retained as the value label for the recoded value of the new variable.
Note: If the source variable has a defined range of user-missing values of the form LO-n, where
n is a positive number, the corresponding user-missing values for the new variable will be
negative numbers.

Visual binning

More Related Content

What's hot (20)

Viewers also liked (20)

Similar to Visual binning (20)

More from Saroj Suwal (20)

Recently uploaded (20)

Visual binning