This book is the only publication that gives complete statistics on the American presidential elections. It not only contains tables for each of the forty-four quadrennial elections from 1789 to 1960, but also has tables for each of the fifty states and each of the eleven historical parties, as well as other tables that feature various interesting sidelights.
In compiling this volume, I have tried to reconcile the numerous contradictory figures that appear in various sources. The original records have been consulted whenever that has been possible. Data filed in the National Archives have been available for the period 1888 to 1960; however, I have had to deal with omissions, inconsistencies, and contradictions. The newspaper files at the Library of Congress have been consulted for the period 1824 to 1888 (for which only a few records are to be found at the Archives), as well as, to a more limited extent, for the period 1892 to 1912.
As a general rule, a newspaper figure was not accepted unless it was corroborated by at least two other papers, and not then when one paper merely quoted another. The perpetuation of typographical errata, incomplete returns, and the misleading figures used by extremely partisan publications has thus been avoided as far as possible. Where data were unobtainable from the National Archives records and the Library of Congress newspaper files, I have used the publications listed in subsequent paragraphs.
One of the reasons for the discrepancies in the figures given by the various authorities is that each candidate for elector does not necessarily receive the same number of votes as his fellow candidates. In the 1940 election, for example, New York had forty-seven electors. The Roosevelt nominees received various votes, some getting 3,251,918, some 3,251,917, others 3,251,- 916, and the rest: 3,251,915. Some of Willkie's candidates received 3,027,478, some 3,027,477, and the rest 3,027,476. The compiler is faced with several alternatives. He can use the high votes, as I have done in all cases where they are determinable; he can use the votes received by each of the higher of the two "at large" candidates for each party, where there were such nominees; he can use the votes received by those whose names appear at the top of the list (these have not necessarily been the "at large" nominees); or he can use the average votes, which would of course produce artificial figures in many instances and would often necessitate the dropping or increasing of ridiculous fractions. Almost all of the states now have the ballots arranged so that the voter ballots for all of the persons listed on the slate of candidates for electors; in other words, there is no "scratching" or possibility of overlooking one or more names.