THOMAS B. ALEXANDER
The Dimensions of Voter Partisan Constancy in
Presidential Elections from 1840 to 1860
By 1840 the nation's second two-party system was fully developed in every region.1 A close, nationwide rivalry between Democrats and Whigs formed the basic configuration of presidential politics, not only in total popular vote but also within most of the states and even within an immense number of counties. Neither party could muster as much as a two-to-one majority in almost three-fourths of the nation's counties in 1840, or even 60 percent in considerably more than one-half.2 This party system was subjected to such intense strain during the succeed-
____________________1
A more general research project on the two-party system of the United States between 1836 and 1880, from which this analysis of continuity is drawn, has been facilitated
by grants from the Social Science Research Council, the University Research Committee
of the University of Alabama, Tuscaloosa, and the University Research Council of the
University of Missouri, Columbia. Computing facility staffs at both universities have
provided essential services. During the development of the data set, of which the
county-aggregate presidential returns are a part, invaluable assistance was furnished by Ross J Cameron, Walter D. Kamphoefner, William W. Beach, Paul E. McAllister, and Harry D. Holmes. Assistance for this specific project was generously furnished by Larry
Steven Demaree and Don Wayne O'Hara, as well as by Walter Kamphoefner. No one
working with presidential election data for this period of United States history should
omit a fervent expression of appreciation to Walter Dean Burnham, whose 1955Presidential Ballots, 1836-1892 opened vistas that quickened the pulse of a generation of
historians before there was a historical archive of machine-readable data at the Inter-
University Consortium for Political and Social Research at Ann Arbor. I am grateful for
the generous assistance provided by Richard P. McCormick.
2
All statements concerning county-level aggregate election data for presidential
elections and all analyses of these data are based on the author's set of county-aggregate
census and election data. This set incorporates presidential, gubernatorial, and congressional election returns for 1838-1878 and 1850, 1860, and 1870 United States census
information. Only the presidential data are used in this essay, except that change in turnout referred to in notes is based, in addition, on population information. The presidential election returns are from Walter Dean Burnham, Presidential Ballots, 1836-1892
(Baltimore, 1955), adjusted or corrected when internal or other evidence justified
changes. The gubernatorial and congressional data were obtained from the Historical
Archive of the I.C.P.S.R. at Ann Arbor.
-70-
