Missouri Census Data Center
Rankster

Data Ranking Utility Application, Vers. 1.01
Section headers are links to the relevant sections of online help.
There is now a video-based tutorial for this application.

Input Data Set:  ##data  .  


I. Choose Variables for Output File(s) and/or Report

Choose (numeric) variable(s) by which to rank
##rankvars -->
Choose ID and other non-ranking variables to keep.
##idother

II. Ranking and Filtering Options

Rank in descending order (largest value gets rank 1)    Rank in ascending order (lowest value gets rank 1)   

The following options (which are optional) only apply if ranking by a single variable:

Limit output to the top/bottom-ranking cases,
and if the value of the ranking variable is more or less than (depending on ascending/descending order) this value:
Check here to keep both highest and lowest ranking cases (ignored if by groups are specified, below).

Optional: Assign ranks within By Groups defined by variable(s)
    Check here if input is already sorted by this/these By Group variable(s).

Optional: Assign group ranks using groups (100 for percentile ranks, 5 for quintile ranks, etc.) Leave blank for normal (ungrouped) ranks.

Option for processing ties: Use low value    Use high value    Use mean rank (e.g. tie for 3rd and 4th gets 3.5)

Optional: Filtering the Data

  1. Limit cases to rank based on minimum universe size, i.e. "filter" (ignore) rows/cases where the universe variable

    (Select: ##selectu )   has a value of at least .   (may not make sense when ranking multiple variables).

  2. Choose observations based on the values of Identifier variables. Select the variable from the drop-down on the left and enter the single value or a list of values (max of 10) separated by colon characters (:) in the box on the right. You can do this for two different Variable / Value combinations. Case matters.
    Variable ##filtvar Value(s)

    Variable ##filtvar2

    Value(s)

III. Output Options

Check if you do NOT want to have a comma-delimited (csv) file generated.
Check if you DO want to produce an output SAS data set (Windows format).

Options for a custom output report:
Report Format:    Standard HTML    Scrollable Table HTML(*)    PDF    None: skip output report.
   * Good choice if using IE browser, but may not work well in others.

Explicitly specify variables to be included in the report.
All variables specified here must also be chosen in Sec I, above. (But not all variables chosen there need to appear in the report.) If you enter nothing here Rankster will assign values based on your choices in Sec I. Entering variable lists here allows you to control the order in which they appear and to be selective about which ones to include. Enter the variable names (spelled exactly as shown in the select lists of Sec. I) in the left-to-right order in which they are to appear in the report. (These name values are not case sensitive.) The report is limited to 12 columns; each varible specified counts as one column except ranking variables, which count as two since both the value and its rank occupy columns. There are 3 sets of variables/report columns here, corresponding to their left-to-right appearance in the report. Specify (by typing their names, separated by blanks):

  1. Identifiers that will identify the row / observation on report lines. You may enter up to 3 such varibles. (If nothing is entered here then the first (up to) 3 character-type non-ranking variables chosen in Sec. I will be used.)

  2. Rank variables. Their values and their ranks will appear in the report. Enter the names of up to 4 variables chosen from the left-side select list in Sec. I. (If nothing is entered here then the first (up to) 4 ranking variables chosen in Sec I will be used.)

  3. Additional variables to be included in the report. Specify up to 9 (subject to overall 12-column limit). Typically, these are numeric variables that are related to the rank variables. (If nothing is entered here then non-ranking, non-character-type variables chosen in Sec.I will be used. The number of such variables will be limited based on the 12-column limit for the report.) To specify that no additional variables are wanted for the report enter a single underscore character here (_).

Check if you want output sorted by the Identifiers (instead of by first Rank variable).

Optionally, you may specify titles and footnotes for the report:

Title1:      
Title2:      
Footnote1:
Footnote2:

Check if you want a statistical summary report of the ranked variables showing mean, median, n and key percentile values.

Internal use only (_debug opt.)


Questions and comments regarding Rankster should be sent via e-mail to John Blodgett.

MCDC Home  ||  OSEDA Home

Generated by the Missouri Census Data Center's prerankster application, revised &revdate.