Last year, the University of Michigan Press published Presidential Elections 1789-2008, a book that shows color maps of the United States for each presidential election, and shows which presidential candidate carried each county. The book is visually beautiful, has 600 pages, and sells for $150. Unfortunately, the authors, Donald R. Deskins, Hanes Walton, and Sherman Puckett, made massive factual errors. The University of Michigan Press is preparing a new edition, but so far, the authors, and officials of the publisher, say they have no interest in correcting any of these errors.

The authors of the book confused several minor parties with each other, probably because all three party names start with the letter “P”. The Peoples Party carried some counties in the presidential elections of both 1904 and 1908, in Georgia, but the book says the Prohibition Party carried these counties. Also, in 1916, the Progressive Party carried Lafourche Parish in Louisiana for president (even though it had no presidential candidate, just candidates for presidential elector who were unpledged for President). But, again, the book says the Prohibition Party carried Lafourche Parish.

Worse than these three somewhat isolated errors are the book’s charts, showing how many votes each presidential candidate received in each state. Even though the University of Michigan houses the Inter-University Consortium that contains the nation’s leading databank of election returns, the authors ignored the University’s data and instead printed incomplete election totals for virtually every minor party and independent presidential candidate they included. The 1904 chart shows that the Socialist Party received 352,270 votes for President, but actually it received 402,714, according to the official returns collected and published in A Statistical History of the American Presidential Elections, the leading authority for presidential vote totals for elections before 1920 and after 1820. The 1908 chart in the University of Michigan book says the Socialist Party received 363,858 votes, but actually it received 420,858. The 1904 chart in the Michigan book says the Prohibition Party only received 61,030 votes, when actually it received 259,163. The 1908 chart in the Michigan book says the Prohibition Party received 21,343, when actually it received 252,704. The 1912 chart in the Michigan book shows the Socialist Party vote at 836,733, when actually it was 901,255.

The book also uses odd standards for deciding which minor party and independent presidential candidates to include in its charts. For 1948 it includes the vote for Strom Thurmond, but not the vote for Henry Wallace, even though Wallace received 1,157,063 votes, only 12,000 fewer than Thurmond. For 1972 the book does not include the vote for John G. Schmitz, who polled 1,105,330 votes. For 1976 it does not include the vote for Eugene McCarthy, who received 756,691 votes. It does not include any Libertarian presidential candidates, not even Ed Clark in 1980 who got 921,299 votes. But it does include Lenora Fulani’s vote in both 1988 and 1992, even though in 1992 she placed sixth (with 73,714 votes), behind not only Ross Perot (who is included) but also behind Andre Marrou (the Libertarian, who received 291,627 votes, but who is not included) and Bo Gritz (the Populist Party nominee, with 107,014 votes, also not included). The book also says that the New Alliance Party was the first party to place its presidential candidate on the ballot of all 50 states, ignoring the fact that the Libertarian Party placed its presidential candidate on the ballot in all states in 1980.

Many of the vote totals for the Democratic and Republican Party nominees are also incorrect. Generally the maps are unaffected by the errors in the charts, and the maps are beautiful and useful. However, even the maps could be more accurate. The authors followed an inconsistent policy on how to handle counties that had not yet been created. For example, Orange County, California, was carved out of Los Angeles County in 1889. The maps for the presidential elections 1852-1876 show Orange County colored the same color as Los Angeles County. This is the best way to handle a county that hadn’t yet been created. But for some reason, the maps for 1880-1888 leave Orange County uncolored, instead of being the color of Los Angeles County. The same inconsistency applies to dozens of other counties, not just that one county.