When preparing the illustrations for my current lecture course on the historical geography of U.S. presidential elections, I was thrilled to discover that Wikipedia’s articles on most of these elections are illustrated with visually striking cartograms of the results, mapped at the county level. In these cartograms, counties are ostensibly scaled to reflect their populations at the time. But as I began to analyze the patterns that they depict, I discovered that they are far from accurate and thus have limited utility.
This realization hit me when I was examining the results of the 1860 election, which resulted in the presidency of Abraham Lincoln – and hence the Civil War. At first glance, the Wikipedia cartogram of this election appeared to be an excellent illustration of the “anybody but Lincoln,” or Fusionist, vote in the greater New York City region. As can be seen in the illustration posted below, New York loomed large in population at the time. It was also distinctly purple, meaning that it gave a majority or plurality of its votes to Fusionist presidential electors.
New York City at the time included only New York County, which is coterminous with the island of Manhattan. Brooklyn, in Kings County, was then a separate city. Queens County, the modern borough of Queens, was at the time largely rural, as was Richmond County, which covers Staten Island. The population imbalances in 1860 were correspondingly considerable, with Manhattan containing over 814,000 residents as opposed to only 57,000 in Queens and 26,000 in Staten Island. But as can be seen from a detail from the Wikipedia cartogram posted below, Queens is portrayed as much larger than Manhattan. Apparently, the counties that constitute the current boroughs of New York City were scaled in accordance with their land areas rather than with their populations at the time, which defeats the purpose of “cartogramic” mapping.
New York City, however, is a special case as it is the only example in the United States of a city that is composed of separate counties. As a result, an error of this sort is perhaps understandable. I therefore decided to test the Wikipedia’s presidential-election cartograms in a different state and in a different election, that of 1916. In this contest, the Democratic candidate, Woodrow Wilson, surprisingly defeated the Republican candidate, Charles Evans Hughes, largely because Wilson eked out an extremely narrow victory in California. The defeat of Hughes in California has generally been attributed to the fact that he snubbed California’s popular Progressive governor, Hiram Johnson, while campaigning in the state. Regardless of this error, Wilson’s victory in California was extraordinary, as the Democratic Party had been almost dead in California only two years earlier. As can be seen in the illustration posted below, in the state’s 1914 gubernatorial election the Democratic candidate received only 12.5% of the vote, with the Republican candidate receiving only 29.3%. The Progressive (“Bull Moose”) candidate, the-once-and-future-Republican Hiram Johnson, won an easy victory with 49.7%.
Owing to California’s massive shift from 1914 to 1916 toward the Democratic Party, I decided that this election was worth analyzing in more detail. I therefore extracted the California portion of the Wikipedia cartogram and expanded it. What I found led me to lose all faith in these Wikipedia visualizations. As can be seen from the figure posted below, the sizes of California’s counties on this cartogram do not represent their populations at the time. This mismatch can be easily seen by comparing the sizes and populations of five counties: San Francisco, San Mateo, Butte, Santa Barbara, and Ventura. The gap between reality and depiction is especially glaring regarding San Francisco County, which is coterminous with the city of San Francisco.
