Measuring Friends-and-Neighbors Voting

Abstract

Although the friends-and-neighbors factor helps explain the results of Democratic primaries in several Southern states, an adequate measure of the phenomenon is conspicuously missing. A procedure for measuring the hometown support, the concentration of that support, and the importance of friends-and-neighbors is developed herein. The 1947 special general election and the 1978 Democratic primaries for the Mississippi U.S. Senate seat are analyzed to show how the measure is better than a previous one, how the importance of friends-and-neighbors in determining election outcomes can be calculated, and how one can detect if the importance is spurious. The analysis demonstrates that friends-and-neighbors was as important in the 1978 first primary as it was in the 1947 special general election. The high importance in the 1978 runoff is partially explained, however, by the emerging North-South cleavage in Mississippi politics.In spite of the conventional wisdom that friends-and-neighbors' vote-i.e., the tendency for &dquo;candidates for state office ... to poll overwhelming majorities in their home counties and to draw heavy support in adjacent counties&dquo; (Key, 1949: 37)-is crucial for understanding politics in several Southern states, little research about the phenomenon has been published. A consequence of this paucity of research is a lack of consensus as to how the effects of friends-and-neighbors can be measured. V. O. Key (1949) identified the presence of the phenomenon by showing that the counties in which a candidate received a high percentage of the vote tended to be those counties that were close to the candidate's hometown; and Earl and Merle Black (1973) used a similar procedure to demonstrate the importance of friends-and-neighbors A UTHORS' NOTE: The authors would like to thank the anonymous reviewers and the editor of the Quarterly for their most helpful suggestions and comments. A special thanks goes to William Gilmore for his assistance in identifying the hometown of R. L. Collins.

Fighting over friends and neighbors: The effect of inter-candidate geographic distance on intra-party competition

“…Key (1949) sparked the interest of many scholars to study the role of localism in elections. What followed was a range of studies looking at various electoral settings, investigating where and for which candidates friends-and-neighbors voting is more prevalent, and discussing how the hometown advantage can be measured correctly (Tatalovich, 1973;Johnston, 1974;Rice and Macht, 1987a;Rice and Macht, 1987b;Van Wingen and Parker, 1979). The degree of urbanization, size of the home county, party affiliation and incumbency status were often moved forward as important factors mediating localism effects.…”

Localism, Factional Fluidity, and Factionalism: Louisiana and Mississippi Gubernatorial Contests

“…In contrast to Allen, Byrd did very well near his hometown and birthplace (the clustering of support) but did not tally well throughout the rest of the state. If one simply regresses his percentage vote with the distance from hometown, as did Tatalovich (1975), one underestimates his hometown and local support (Van Wingen and Parker, 1979). When friends and neighbors' support is present, the relationship between percentage vote and distance from hometown is nonlinear.…”

“…In an earlier work (Van Wingen and Parker, 1979), Parker and I developed a measure that identified when the vote disproportionately clusters around the candidate's home. Rather than following Tatalovich's procedure, we regressed the log of the percentage vote with the log of the distance from the candidate's hometown.…”

“…Two procedures are available for calculating the percentage variance in a candidate's vote that localism explains, after controlling for regionalism. Johnston (1973) developed one; Parker and I (Van Wingen and Parker, 1979), the other, both examined the residuals to see whether they conformed to a friends and neighbors' pattern. Johnston, however, analyzed the residuals to see if friends and neighbors' effect can be observed after controls.…”

“…Key's classic has, of course, spawned much research into the voting patterns of the South. Several articles examine his notions on the causes of factionalism (see, in particular, Cannon, 1978Cannon, , 1980Echols and Ranney, 1976), and a number of works investigate the phenomenon of friends and neighbors either tangentially (Black andBlack, 1973a, 1973b) or more directly (Cotter, 1982;Tatalovich, 1975;Van Wingen and Parker, 1979). The relationship between the three distinct measures, however, has received little attention.…”