Counting and Recounting: 1924 Detroit and the Limitations of Electoral Data

Russell Fehr, Independent Scholar

In the field of urban politics, election results are frequently used as a major source of data concerning the opinions of the general public, and have been used to demonstrate racial, ethnic, religious, and class divisions among the electorate. This use of electoral data, however, must always be done with caution for various reasons. One of the most significant dangers is that this use of data presumes that the results as reported are both accurate and unquestionable. A case study for the limitations of electoral data can be found in the 1924 Detroit mayoral election, which ultimately hinged on the counting of around 15,000 votes. Charles Bowles, who ran as a write-in candidate with the backing of the Ku Klux Klan, argued for a broad interpretation of write-in votes in which all ballots that could be interpreted as being for him would be counted. The City Election Commission, comprised of Detroit elected officials, engaged in a strict interpretation of counting write-in votes, leading to Bowles losing by 14,000 votes. This dispute over counting the ballots can be situated in two different local contexts. In previous decades, the fairness of the vote in Detroit had been a subject of substantial dispute, leading to major changes in local election law. Similarly, Bowles, as the candidate of the Ku Klux Klan, was perceived as a threat to the political and social order of Detroit, with a Klan riot during the campaign for mayor making this threat even starker. Overall, this dispute demonstrates limitations to electoral data by showing how electoral data is both less precise and more heavily contested than is often acknowledged. It suggests that future use of electoral data should be more cautious in nature, acknowledging that full precision may not be possible with this data.

No extended abstract or paper available

 Presented in Session 73. Methods of Inquiry