Bulletins
The statute of censorship that formed the Ohio Board of Film Censorship declared that the censors should approve any film that was "of a moral, educational, or amusing and harmless character" (Carmen 11). That statement is intentionally vague and in positive language, it does not tell us what the censorship board routinely did not approve of in the films submitted. For the censorship process every film submitted was categorized as approved, approved with eliminations, or rejected. As shown in the Overview, the rejected films make up the smallest percentage of films censored and, more importantly, not much information was given when a film was rejected. The majority of the time the censors just wrote "on account of being harmful." Because of the quantity of films approved with eliminations is greater, and the general specificity of the elimination data, this research focuses heavily on the bulletins. For a study of rejected films see the Case Studies section.
The bulletins are summaries of the eliminations requested from films that are deemed "approved with eliminations." A bulletin was created each week that collected together all of the eliminations requested of every submitted film that fit that category. The bulletin would list the title of film, number of reels, and the eliminations: cut out subtitle, cut out scene where woman undresses, cut out men smoking, etc. The corpus of this research was gathered from the archive at Ohio History Connection and consists of the bulletins: September 1915 - February 1916, 1924, 1925, 1929, 1933, 1934, 1940, 1944, 1947, 1952, 1953, 1954. These dates were selected specifically for their historical context, both in terms of film history and general U.S. history.
The bulletins are analyzed using a text analysis software called Voyant, which organizes large quantities of data so that it can be examined and researched with specific question. It is important to note that Voyant Tools does not do any analysis of its own, it merely organizes and presents the data in various ways for the researcher. As can be seen in the Censorship Topics section, these tools are an effective way to explore a large corpus of information. This collection of digital tools allows the researcher to examine a large corpus that would be extremely time consuming, and subject to error, if done traditionally. In order to get the bulletins into the right condition for the Voyant research, the documents had to go through several transitions. First, pictures were taken of the documents in the archive at Ohio History Connection in Columbus, OH. Then those images were scanned for Optical Character Recognition (OCR) with a program called Abbyy FineReader, which allows for the researcher to correct the OCR process as it occurs, thus allowing the program to learn how to better read the documents. This was effective, and the program made short work of the bulletins. However, the bulletins are form documents, meaning they contain a lot of information that was not necessary for the text analysis of eliminations. So the OCR'd documents were stripped down into plain text, containing only the eliminations requested. The bulletins were weekly documents, but in order to make the corpus manageable they were grouped together by year and uploaded into Voyant as the corpus for the bulletins.
The corpus of this research is quite large (in a Word document it is over 600 pages). Because of this size it was not feasible to correct the OCR process as it occurred, and so the data does contain errors. These errors involve misspellings and a small number of duplicate entries. These errors are due to imperfections in the manuscripts and the limitations of OCR software. While having data with errors is not ideal, these errors are not significant enough to derail the text analysis process. As will be shown in this section, the quantity of the data means that trends and conclusions can still be drawn from the data. Additionally it is important to note that the six months of bulletins from 1915-1916 were OCR'd clean, so the data for the close reading is free from errors. This is one of the reasons that the results from that analysis provides part of the framework for the larger censorship topics.
It is recommended to go through the analysis of bulletins in the order presented, for the first section establishes a foundation of the research and explains the various tools being used. The first section of the text analysis is a combination of Voyant tools and Close Reading of the six months of bulletins for 1915-1916. The second section is one of the major components of the research, the analysis of the larger corpus to determine the Censorship Topics, using Voyant to analyze the Board's bulletins to determine the main topics of censorship and how those topics changed over time. The final section, Specific Examples, reinforces the findings of the first two while answering some tangential but engaging questions.