Overview (MULT RESPONSE command)

MULT RESPONSE displays frequencies and optional percentages for multiple-response items in univariate tables and multivariate crosstabulations. Another procedure that analyzes multiple-response items is TABLES, which has most, but not all, of the capabilities of MULT RESPONSE. TABLES has special formatting capabilities that make it useful for presentations.

Multiple-response items are questions that can have more than one value for each case. For example, the respondent may have been asked to circle all magazines read within the last month in a list of magazines. You can organize multiple-response data in one of two ways for use in the program. For each possible response, you can create a variable that can have one of two values, such as 1 for no and 2 for yes; this is the multiple-dichotomy method. Alternatively, you can estimate the maximum number of possible answers from a respondent and create that number of variables, each of which can have a value representing one of the possible answers, such as 1 for Time, 2 for Newsweek, and 3 for PC Week. If an individual did not give the maximum number of answers, the extra variables receive a missing-value code. This is the multiple-response or multiple-category method of coding answers.

To analyze the data entered by either method, you combine variables into groups. The technique depends on whether you have defined multiple-dichotomy or multiple-response variables. When you create a multiple-dichotomy group, each component variable with at least one yes value across cases becomes a category of the group variable. When you create a multiple-response group, each value becomes a category and the program calculates the frequency for a particular value by adding the frequencies of all component variables with that value. Both multiple-dichotomy and multiple-response groups can be crosstabulated with other variables in MULT RESPONSE.

Options

Cell Counts and Percentages. By default, crosstabulations include only counts and no percentages. You can request row, column, and total table percentages using the CELLS subcommand. You can also base percentages on responses instead of respondents using BASE.

Format. You can suppress the display of value labels and request condensed format for frequency tables using the FORMAT subcommand.

Basic Specification

The subcommands required for the basic specification fall into two groups: GROUPS and VARIABLES name the elements to be included in the analysis; FREQUENCIES and TABLES specify the type of table display to be used for tabulation. The basic specification requires at least one subcommand from each group:

  • GROUPS defines groups of multiple-response items to be analyzed and specifies how the component variables will be combined.
  • VARIABLES identifies all individual variables to be analyzed.
  • FREQUENCIES requests frequency tables for the groups and/or individual variables specified on GROUPS and VARIABLES.
  • TABLES requests crosstabulations of groups and/or individual variables specified on GROUPS and VARIABLES.

Subcommand Order

  • The basic subcommands must be used in the following order: GROUPS, VARIABLES, FREQUENCIES, and TABLES. Only one set of basic subcommands can be specified.
  • All basic subcommands must precede all optional subcommands. Optional subcommands can be used in any order.

Operations

  • Empty categories are not displayed in either frequency tables or crosstabulations.
  • If you define a multiple-response group with a very wide range, the tables require substantial amounts of workspace. If the component variables are sparsely distributed, you should recode them to minimize the workspace required.
  • MULT RESPONSE stores category labels in the workspace. If there is insufficient space to store the labels after the tables are built, the labels are not displayed.

Limitations

  • The component variables must have integer values. Non-integer values are truncated.
  • A maximum of 100 existing variables named or implied by GROUPS and VARIABLES together.
  • A maximum of 20 groups defined on GROUPS.
  • A maximum of 32,767 categories for a multiple-response group or an individual variable.
  • A maximum of 10 table lists on TABLES.
  • A maximum of 5 dimensions per table.
  • A maximum of 100 groups and variables named or implied on FREQUENCIES and TABLES together.
  • A maximum of 200 non-empty rows and 200 non-empty columns in a single table.