Preliminary Analysis of the 1968 Wallace Vote in the Southeast

PRELIMINARY ANALYSIS OF THE 1968 WALLACE VOTE IN THE SOUTHEAST Stephen S. Birdsall* During the past century, national elections in the United States have rarely included a significant showing by a third party. Political parties other than the Republican and the Democratic have usually suffered from limited constituencies, either regionally or in terms of issues. An extremely interesting phenomenon, therefore, was the presence of a “third” party in 1968. Before the Fall elections, it was believed to have had considerable appeal across the country. George Wallace’s American Independent Party (or occasionally, American Party or Independent Party) was frequently discussed as having the best opportunity since Theodore Roosevelt’s efforts in 1912 of upsetting the political plans of the two larger parties. The general results of the 1968 Presidential Election are known. Although receiving the electoral support of nearly 10 million voters, Wallace polled over half of this total in eleven Southern states. Furthermore, the American Independent Party carried only five states, all in the Southeast: Arkansas, Louisiana, Mississippi, Alabama, and Georgia. The 1968 Wallace-for-President effort, therefore, has some appearances of the limited regional appeal typical of third-party activities in Presidential politics. (1) Less clear than the general results are the characteristics and spatial pattern of elector support. This paper reports briefly upon preliminary research which deals with voter characteristics and the distribution of voter support for George Wallace in 1968. Data have been analyzed by statistical methods for eight states in the Southeast: North and South Carolina, Ten­ nessee, Georgia, Florida, Alabama, Mississippi, and Louisiana. Although Wallace received many votes in Arkansas, Virginia, Texas, and Kentucky, among other states, data for one of the variables considered important to the analysis were not available for these states. (2) Visual examination of the election results is extended to include Virginia and Arkansas. ORGANIZATION AND LIMITATIONS. The analyses presented here have been divided into three sections. First, the county-to-county distribution of Wallace vote is examined. This pattern is the per cent of votes received by Wallace from the total number cast in each county. Second, an attempt is made to identify some of the significant characteristics ofWallace supporters. This attempt is made through statistical analysis of county population *Dr. Birdsall Is assistant professor of geography at the University of North Carolina, Chapel H ill. The paper was accepted for publication in August 1969. 56 So u t h ea st er n G e o g r a ph er characteristics for the eight-state region. Third, the relationship between degree of support and the population providing this support is studied. As location is taken to be a significant aspect of this relationship, the purpose of this last set of analyses is to determine whether or not there is significant regional variation in the relationship between support for Wallace and characteristics of the population giving it. More loosely, is the appeal of the American Independent Party interpreted differently by voters in one portion of the Southeast than by voters elsewhere in the region? Or again, is the Southeast homogeneous in the interpretation of the appeal from a largely regional candidate? The third portion of this paper, therefore, is a combina­ tion of the analyses of the first two sections. There are numerous difficulties attending the analyses. As with any study using general population characteristics and voting returns, it is impossible to determine the character of the non-voting population. It should also be made clear that the use of grouped (in this case, county) data precludes the assignment of motive or detailed characteristics to any member of the group. Therefore, although conclusions can be stated in terms of specific bases for voter support, individuals cannot be characterized with these results. (3) Additional problems are raised by the use of voter registration records. For example, some counties have more registered voters than eligible voters; deceased and departed registrants might not have been purged from the rolls. More rigor is lost by the necessity of examining 1968 election returns in light of 1960 population characteristics. In spite of these and other problems, the conclusions reached are believed valid as indicators of the general relation between the phenomena studied as...