Overview (RANK command)

RANK produces new variables containing ranks, normal scores, and Savage and related scores for numeric variables.

Options

Methods. You can rank variables in ascending or descending order by specifying A or D on the VARIABLES subcommand. You can compute different rank functions and also name the new variables using the function subcommands. You can specify the method for handling ties on the TIES subcommand, and you can specify how the proportion estimate is computed for the NORMAL and PROPORTIONAL functions on the FRACTION subcommand.

Format. You can suppress the display of the summary table that lists the ranked variables and their associated new variables in the active dataset using the PRINT subcommand.

Basic Specification

The basic specification is VARIABLES and at least one variable from the active dataset. By default, the ranking function is RANK. Direction is ascending, and ties are handled by assigning the mean rank to tied values. A summary table that lists the ranked variables and the new variables into which computed ranks have been stored is displayed.

Subcommand Order

• VARIABLES must be specified first.
• The remaining subcommands can be specified in any order.

Operations

• RANK does not change the way the active dataset is sorted.
• If new variable names are not specified with the INTO keyword on the function subcommand, RANK creates default names. (Note: The automatically generated new variable names are limited to a maximum length of 8 bytes.)
• RANK automatically assigns variable labels to the new variables. The labels identify the source variables. For example, the label for a new variable with the default name RSALARY is RANK of SALARY.