Courtney Brown, Ph.D.
A Brief Biography
Curriculum Vitae
Career Guidance Videos
Political Music Videos
Political Music Articles
R Tutorial Videos
Data Sets and
Computer Programs
Speculative Nonfiction
The Farsight Institute
Book Reviews
Publicity Photos
Speaking Requests
Farsight Prime
African Television
Music Videos

Follow on Facebook Courtney Brown
Follow on FB
Follow on Twitter Courtney Brown
Follow on Twitter
Courtney Brown on Instagram


Variable Descriptions for U.S. Aggregate Data Sets

The aggregate county-level data sets for the United States which I have used in many of my published research were originally supplied to me by the Interuniversity Consortium for Social and Political Research (ICPSR). The data were originally in hundreds of separate files containing political and census data which were sorted by year and state. I had to assemble single data sets for specific periods of time from these separate files. To avoid having to write variable labels for all of the variables, I used a common scheme for the variable names. This scheme is described below.

The variables for political data generally have names with eight characters. The first character indicates whether the variable is for presidential or congressional data (P or C, respectively). The next four characters indicate either the party or if the data are the total vote, where

0100 = Democratic Party vote (county-level)
0200 = Republican County vote (county-level)
TOTA = Total Vote (county-level)

Following the party or total vote designators is the year for the data.

For example, the variable P0100920 would be the county-level presidential vote in the U.S. for the Democratic Party in the year 1920. Similarly, C0200932 would be the county-level congressional vote in the U.S. for the Republican Party in the year 1932.

Eligible voters (as determined by age using U.S. Census data) are found in variables beginning with the letter "E" followed by the year. Thus, E1930 represents the total adults 21 years and older as determined by the U.S. Census for the year 1930.

Other variables should be self-explanatory, where the meaning either is explicit in the name of the variable or the meaning is stated in its variable label.

For the mcp5084 data set (the "mcp" stands for merged census and political data), the data are coded somewhat differently. In this instance single letters are used to designate the party, and two digits are used for the year. Thus, PD52 stands for county-level presidential votes for the Democratic Party in the year 1952. Similary, CR80 would be congressional votes for the Republican Party in the year 1980.

For all of the data sets, the variable COUNTY is the identifier for each county, which begins with the two-digit state number identifier used by the ICPSR followed by the county number used by the ICPSR. The exact SAS statement that was used to create this variable is: