Description
|
This corpus contains one week of election coverage by the five largest—by circulation—newspapers in the state of Texas—Dallas Morning News, Houston Chronicle, San Antonio Express-News, Fort Worth Star-Telegram, Austin American-Statesman. The period of coverage is from October 10, 2018, to October 16, 2018, the first week after the Texas voter registration deadline. I collected via two methods. I searched the NewsBank news archives for the Houston Chronicle, San Antonio Express-News, Fort Worth Star-Telegram, and Austin American-Statesman. The archival holdings for the Dallas Morning News do not progress past 2016. I conducted the searches within specific sources using the Boolean search input ("campaign" or "candidate" or "election"). To verify the completeness of the sample I performed the same search with the addition of the term “Beto.” As the most popular candidate with a name unlikely to provide false returns, “Beto” was a strong robustness test for the search phrase. No additional articles were returned for the first source, The Austin American-Statesman, so the search input was accepted. I collected the Dallas Morning News sample from the EBSCOhost Newspaper Source Plus. The initial search returned 253 articles across the five sources. I removed duplicates from within papers—but not across papers. Because the analysis for which I collected the corpus focuses on news norms when discussing individual candidates, I removed all articles not containing a discussion of at least one specific candidate. The final corpus contains 126 articles. Paragraphs are enumerated. ---- The dataset was expanded in December 2021 to include a full second week of coverage (October 31, 2018, to November 6, 2018). Details on that expansion are below. This corpus includes two weeks of election coverage by the five largest—by circulation —newspapers in the state of Texas. The first period of coverage is from October 10, 2018, to October 16, 2018, the first week after the Texas voter registration deadline. The second is from October 31, 2018, to November 6, 2018, the week leading up to election day, which includes the last days of early voting. The five largest papers in the state by circulation are the Dallas Morning News, Houston Chronicle, San Antonio Express-News, Fort Worth Star-Telegram, Austin American-Statesman (”Circulation of most popular Texas newspapers in the U.S. 2016” n.d). The corpus for analysis consists of two full weeks of election coverage for each of the listed sources. I collected these via searches of EBSCOhost’s Newspaper Source Plus and NewsBank’s news archives. I conducted the searches within specific sources using the Boolean search input (“campaign” or “candidate” or “election”). To verify completeness of sample, I performed the same search with the addition of the term “Beto.” As the most popular candidate with a name unlikely to provide false returns, “Beto” was a strong robustness test for the search phrase. No additional articles were returned for the first source, The Austin American-Statesman, so the search input was accepted. The initial search returned 253 articles for the first week and 359 for the second. I removed duplicates from within papers—but not across papers. Because the analysis focuses on the gendered selection of candidates and sources, I removed all articles not containing a discussion of at least one specific candidate (e.g., articles discussing Heisman Trophy “candidates” or lists of polling locations). This led to a final sample of 126 articles for the first week and 159 for the second, 285 total. (2021-01-20)
|