Rank Data Task

Example: Ranking Students by Height within Age

About the Rank Data Task

The Rank Data task computes ranks for one or more numeric variables across the rows in a table and includes the ranks in an output table.

An example of when you might use the Rank Data task is to rank product sales. In this case, the ranking variable would show the order of product sales. The product with the highest number of sales would be ranked first.

Example: Ranking Students by Height within Age

In this example, you want to rank the students in your class by age and height.

To create this example:

In the Tasks section, expand the Data folder and double-click Rank Data. The user interface for the Rank Data task opens.
On the Data tab, select the SASHELP.CLASS data set.

Assign columns to these roles:

Role	Column Name
Roles
Columns to rank	Height
Additional Roles
Rank by	Age

Open the Options tab. From the Rank order drop-down list, select Largest to smallest.
To run the task, click .

The Rank Data task creates an output data set. In SAS Studio, this data set opens on the Output Data tab. This data set contains the additional rank_Height column, which shows where that student ranks within her age group. For example, in the 11-year-old age group, Joyce is ranked number 2. In the 12-year-old age group, Louise is ranked number 5.

Output Data Set Created by the Rank Data Task

Assigning Data to Roles

To run the Rank Data task, you must assign a column to the Columns to rank role.

Role	Description
Roles
Columns to rank	Each column that is assigned to this role is ranked. You must assign at least one variable to this role. By default, the rankings column is given the name rank_column-name, where column-name is the name of the original column.
Additional Roles
Rank by	When you assign one or more columns to this role, the input table is sorted by the selected column or columns and rankings are calculated within each group.
Output Data Set
Create new variables for the ranked variables	specifies that the output table contains the original columns as well as the ranked columns. If you want to replace the original column with the ranked columns, clear the Create new variables for the ranked variables check box. By default, the ranked column is given the name rank_column-name, where column-name is the name of the original column.
Show output data	specifies whether to display all or a subset of the output data in the results.

Setting Options

You must select at least one output option.

Option Name	Description
Options
Ranking method	specifies the method to use when ranking the data. Here are the valid values: Ranks partitions the original values into 100 groups, in which the smallest values receive a percentile value of 0 and the largest values receive a percentile value of 99. Quantiles partitions the original values into one of these quantiles. Percentiles partitions the data into 100 groups, in which the smallest values receive a percentage value of 0 and the largest values receive a percentage value of 99. Deciles partitions the original values into 10 groups, in which the smallest values receive a decile value of 0 and the largest values receive a decile value of 9. Quartiles partitions the original values into four groups, in which the smallest values receive a quartile value of 0 and the largest values receive a quartile value of 3. N-tile groups partitions the original values into n groups, in which the smallest values receive a value of 0 and the largest values receive a value of n–1. Specify the value of n in the Number of groups box.
Ranking method (continued)	Fractional ranks computes the fractional ranks by using either a denominator of N or N+1. A denominator of N computes fractional ranks by dividing each rank by the number of observations that have nonmissing values of the ranking variable. A denominator of N+1 computes fractional ranks by dividing each rank by the denominator n+1, where n is the number of observations that have nonmissing values of the ranking variable. Percentages divides each rank by the number of observations that have nonmissing values of the variable and multiplies the result by 100 to get a percentage.
Ranking method (continued)	Normal scores of ranks computes normal scores from the ranks. The resulting variables appear normally distributed. Here are the formulas: Blom formula $y sub i , equals , cap phi super negative 1 end super , open . fraction open , r sub i , minus , 3 eighths , close , over open n plus , 1 fourth , close end fraction . close. Click image for alternative formats.$ Tukey formula $y sub i , equals , cap phi super negative 1 end super , open . fraction open , r sub i , minus , 1 third , close , over open n plus , 1 third , close end fraction . close. Click image for alternative formats.$ van der Waerden $y sub i , equals , cap phi super negative 1 end super , open . fraction r sub i , over open n plus 1 close end fraction . close. Click image for alternative formats.$ In these formulas, is the inverse cumulative normal (PROBIT) function, r_i is the rank of the ith observation, and n is the number of nonmissing observations for the ranking variable. Note: If you set the If values are tied, use option, the Rank Data task computes the normal score from the ranks based on non-tied values and applies the ties specification to the resulting score. Savage scores of ranks computes Savage (or exponential) scores from the ranks. Note: If you set the If values are tied, use option, the Rank Data task computes the Savage score from the ranks based on non-tied values and applies the ties specification to the resulting score.
If values are tied, use	specifies how to compute normal scores or ranks for tied data values. Default method assigns the default method for your ranking method. If you select Percentages or Fractional ranks as the ranking method, the high value is the default. For all other ranking methods, the mean is the default. Mean of ranks assigns the mean of the corresponding rank or normal scores. High rank assigns the largest of the corresponding ranks or normal scores. Low rank assigns the smallest of the corresponding ranks or normal scores. Dense rank (ties are the same rank) computes scores and ranks by treating tied values as a single-order statistic. For the default method, ranks are consecutive integers that begin with the number 1 and end with the number of unique, nonmissing values of the variable that is being ranked. Tied values are assigned the same rank.
Rank order	specifies whether to list the values from smallest to largest or from largest to smallest.