Selected Papers of #AoIR2023: The 24th Annual Conference of the Association of Internet Researchers Philadelphia, PA, USA / 18-21 Oct 2023 Suggested Citation (APA): Ringel,S., and Davidson, R. (2023, October). Demographic, Occupational and Professional Predictors of Tweet Deletion Among U.S. Journalists. Paper presented at AoIR2023: The 24th Annual Conference of the Association of Internet Researchers. Philadelphia, PA, USA: AoIR. Retrieved from http://spir.aoir.org. DEMOGRAPHIC, OCCUPATIONAL AND PROFESSIONAL PREDICTORS OF TWEET DELETION AMONG U.S. JOURNALISTS Sharon Ringel Department of Communication, University of Haifa Roei Davidson Department of Communication, University of Haifa Introduction. Twitter and other social media platforms have become a key infrastructure for contemporary journalism (Kleis Nielsen & Ganter, 2018). Journalists use Twitter to gather information (Powers & Vera-Zambrano, 2017). They employ it to maintain a relationship with their audiences and sources and through that to cultivate their individual personal and professional brand as well as their employer’s brand (Molyneux et al., 2018). They also communicate with journalistic peers and hence participate in the ongoing production of aN interpretive community (Carlson, 2016). While most research has focused on journalists’ publication practices, much less work has examined their decision to delete content they have already published on social media, an action that could have a meaningful impact on twitter as a repository of public knowledge. In in- depth interviews, American journalists reported deleting tweets frequently. They also claimed such practices were common among their peers (Ringel & Davidson, 2022). In this study, we examine how journalists’ demographic identity, occupational status and professional standing might be related to their tendency to delete tweets. Occupational and organizational attributes of journalism work. Many workers in the culture industries experience considerable precarity: they often work on short-term contracts, constantly in search of the next gig with minimal job security (Gill & Pratt, 2008; Hesmondhalgh & Baker, 2009; Sperlich, 2011). Journalists are no exception: many journalists - especially those at an early stage of their career - adopt entrepreneurial labor practices out of choice or necessity (Davidson & Meyers, 2015). In such an employment environment, journalists (like many other workers) must carefully craft an employable public image that will appeal to many prospective employers. In interviews with American journalists, they linked their feeling of occupational uncertainty with their decision to delete tweets as a means of constructing an online presence that would not offend employers and therefore would not endanger their future employment prospects (Ringel & Davidson, 2022). Journalists’ harassment: its demographic and occupational dimensions. As they maintain their online presence on Twitter, journalists around the world are encountering various forms of online harassment with women and ethnic minorities more likely to be targeted (Waisbord, 2022). Survey data suggests most American journalists experience some form of online harassment. However, women (and especially young women) as well as those who are personally visible when they do their journalistic work reported experiencing more harassment than others (Lewis et al., 2020). Female journalists in five countries reported facing extensive harassment, with visible journalists such as those working as television reporters exposed to more harassment. They also suggested that minority female journalists were even likelier to be exposed to hateful and demeaning treatment online (Chen et al., 2020). Harassment has consequences for the ways journalists practice journalism and engage with their audience. When journalists reported experiencing harassment they perceived the audience as less rational and more different than themselves and evaluated their interaction with the audience as of a lower quality. Such perceptions might weaken their commitment to provide the audience future access to their past tweets. Slightly less than 50% of respondents claimed they changed what they posted in response to harassment (women more than men) and approximately 8% reported deactivating or deleting their account (Lewis et al., 2020). Some interviewed female journalists who experienced harassment said that they limited the topics they covered and toned down the critical tone they took. Some also reported deleting abusive online comments either manually or automatically using tools to block abuse (Chen et al., 2020). Such findings suggest that those journalists especially targeted by harassment might be more likely to carefully curate their online presence and hence might delete more tweets. American journalists we interviewed argued harassment - especially of female and minority journalists - led those journalists to delete tweets whether selectively or en masse (Ringel & Davidson, 2022). Research questions. Given existing evidence that the demographic characteristics of journalists, and their professional and occupational standing impact the occupational and personal risks they face as they maintain their online presence, we wish to track the deletion practices of a large sample of journalists more directly to examine whether and how individual-level characteristics of the journalists are related to their tendency to delete tweets they have previously published. We, therefore, ask: RQ1: Are demographic characteristics (ethnic identity, gender, age, educational attainment) of American journalists related to their likelihood of deleting tweets? RQ2: Is the occupational status of American journalists as free-lancers related to their likelihood of deleting tweets? We will also investigate whether their professional characteristics are related to their tendency to delete tweets. We will more specifically ask: RQ3a: Are a journalist’s years of experience related to their likelihood of deleting tweets? RQ3b: Is the visibility of a journalists’ role related to their likelihood of deleting tweets? RQ3c: Is professional journalistic training related to their likelihood of deleting tweets? Methods. We adapted the Politwoops tool used to track and display the deleted tweets of American politicians (Willis, 2023) to track the tweeting and deletion practices of a diverse cluster sample of approximately 3,370 employees of news organizations who maintain personal accounts on Twitter over a period of approximately 3 months in the fall of 2022 (August 17 to November 22). We sampled accounts of employees working in 97 major American news organizations that had more than 10 million unique digital users in late 2020 (“State of the News Media Methodology,” 2021). In addition, we sampled journalists working for Fox News and those identified by Cision (a commercial media relationship management database) as freelance journalists. From each sampled news organization we randomly selected 30 individuals listed as affiliated with the organization. We excluded from analysis accounts that had fewer than 300 followers and those accounts identified with individuals who though affiliated with the organization did not have a journalistic role within it (e.g., marketing, advertising, administration). From each organization we identified a maximum of 20 journalists. We used identical criteria to randomly sample 430 journalists identified on Cision as free-lancers. The Politwoops tool allowed us to register each instance of deletion as well as the tweets that were not deleted. We will use a textual similarity detection algorithm (Briggs, 2021) to compare each deleted tweet to the subsequent five tweets an account posted to identify cases in which deletion was not substantive but rather used as a means of lightly editing a tweet (given the absence of an edit button). We will measure likelihood of deletion as the proportion of all substantive deletions to the overall number of tweets published during the data collection period. We will combine this data with ongoing manual content analysis of their Twitter profile, their Linkedin profile, and the Cision database to code demographic characteristics (ethnic identity, gender, year they began their undergraduate studies [as proxy of age], educational attainment), occupational status (free-lancer [yes/no]), and professional characteristics (years of journalistic experience, journalistic role, journalism training [yes/no]). As the content analysis proceeds, we are assessing coding reliability. To examine the relationship of these variables with deletion likelihood we will employ a multivariate regression analysis. With deletion data already collected and content analysis near completion we will be able to present a full analysis of the data at the conference. References Briggs, J. (2021, September 2). BERT For Measuring Text Similarity. Medium. https://towardsdatascience.com/bert-for-measuring-text-similarity-eec91c6bf9e1 Carlson, M. (2016). Metajournalistic Discourse and the Meanings of Journalism: Definitional Control, Boundary Work, and Legitimation: Metajournalistic Discourse and the Meanings of Journalism. Communication Theory, 26(4), 349–368. https://doi.org/10.1111/comt.12088 Chen, G. M., Pain, P., Chen, V. Y., Mekelburg, M., Springer, N., & Troger, F. (2020). ‘You really have to have a thick skin’: A cross-cultural perspective on how online harassment influences female journalists. Journalism, 21(7), 877–895. https://doi.org/10.1177/1464884918768500 Davidson, R., & Meyers, O. (2015). Toward a Typology of Journalism Careers: Conceptualizing Israeli Journalists’ Occupational Trajectories. Communication, Culture & Critique, 9(2), 193–211. Gill, R., & Pratt, A. (2008). In the Social Factory?: Immaterial Labour, Precariousness and Cultural Work. Theory, Culture & Society, 25(7–8), 1–30. https://doi.org/10.1177/0263276408097794 Hesmondhalgh, D., & Baker, S. (2009). “A very complicated version of freedom”: Conditions and experiences of creative labour in three cultural industries. Poetics, 38(1), 4–20. https://doi.org/10.1016/j.poetic.2009.10.001 Kleis Nielsen, R., & Ganter, S. A. (2018). Dealing with digital intermediaries: A case study of the relations between publishers and platforms. New Media & Society, 20(4), 1600–1617. Lewis, S. C., Zamith, R., & Coddington, M. (2020). Online Harassment and Its Implications for the Journalist–Audience Relationship. Digital Journalism, 8(8), 1047–1067. https://doi.org/10.1080/21670811.2020.1811743 Molyneux, L., Holton, A., & Lewis, S. C. (2018). How journalists engage in branding on Twitter: Individual, organizational, and institutional levels. Information, Communication & Society, 21(10), 1386–1401. https://doi.org/10.1080/1369118X.2017.1314532 Powers, M., & Vera-Zambrano, S. (2017). How journalists use social media in France and the United States: Analyzing technology use across journalistic fields. New Media & Society, 20(8), 2728–2744. https://doi.org/10.1177/1461444817731566 Ringel, S., & Davidson, R. (2022). Proactive ephemerality: How journalists use automated and manual tweet deletion to minimize risk and its consequences for social media as a public archive. New Media & Society, 24(5), 1216–1233. https://doi.org/10.1177/1461444820972389 Sperlich, R. (2011). The mixed blessing of autonomy in digital cultural production: A study on filmmaking, press photography and architecture in Austria. European Journal of Communication, 26(2), 133–146. https://doi.org/10.1177/0267323111401611 State of the News Media methodology. (2021, July 27). Pew Research Center’s Journalism Project. https://www.pewresearch.org/journalism/2021/07/27/state-of-the-news-media- methodology/ Waisbord, S. (2022). Online Trolling of Journalists. In S. Allan (Ed.), The Routledge Companion to News and Journalism (2nd ed., pp. 149–158). Routledge. https://doi.org/10.4324/9781003174790-18 Willis, D. (2023, February 24). After a Decade of Tracking Politicians’ Deleted Tweets, Politwoops Is No More. ProPublica. https://www.propublica.org/article/politwoops-deleted- tweets-twitter-politicians-musk