Computational Journalism and Local News

The following provides an in-depth summary of current work on computational journalism and local news.

The rise of digital media platforms transformed the way that news is produced and distributed, especially with regards to local news. Simultaneously, digital media platforms transformed the way in which news media organizations engage with their public audiences. This transformation is particularly salient in the context of local news, where circulation declines paired with struggles to gain online audiences have created a myriad of challenges. Compared to other news entities, local newspapers face particularly tight constraints on resources. National and regional newspapers, as well as online news sources, are increasingly reliant on algorithmic technology as a means for helping to produce, curate and distribute news; in addition, algorithmic technology is helping journalists to produce articles faster and more efficiently (Diakopoulos & Koliska, 2016; Thurman et al., 2016).

Algorithmic technology, in the context of newspapers, includes technology such as mobile applications that automatically serve articles to consumers, search tools that help reporters to locate sources, and computer programs that help to write articles. While much has been made of the potential of algorithms to help advance journalism (Coddington, 2015), there have been numerous calls for increased transparency to better understand how journalism is being automated through algorithms (Diakopoulos & Koliska, 2016). Algorithms have the potential to create significant efficiencies with regards to the production of journalistic content, and further have the potential to improve the way content is delivered to audiences. These efficiencies are particularly appealing at the local level, where there is a clear desire to develop new models of business and production.

Thus, this work differs from prior studies of algorithms and news production by specifically interrogating the local context of algorithmic technology. Local news organizations face a unique set of challenges with regards to producing and distributing news through the use of algorithmic technology. Local news organizations face severe constraints on resources; the smaller scale of most local news organizations means that fewer resources are devoted to the development of new technology. In part, leaders within local newsrooms often find that their job roles are ambiguous in nature and cross function boundaries (e.g. blending editor with marketing manager) (Swasy, 2016; Young & Carson, 2016). The implementation of algorithmic technology is therefore complicated at the local level by immediate business challenges and immense competition from larger scale efforts.

Despite challenges brought on by algorithmic technology, local news organizations continue to serve an important role providing news and information to their communities. Local news organizations provide a critical source of information for an informed citizenry, and help to support a rich local political ecosystem and democratic decision making process (Downie & Schudson, 2009). Yet at the local level increasing competition for audiences from national newspapers, social media, and other information sources has meant that most local news organizations have been forced to experiment with new business models, but without sufficient resources (White, Pennycook, Perrin, & Hartley, 2017). As a result of the proliferation of information sources, increased competition for advertising dollars, and a host of other factors, local newspapers are in a marked era of decline (Nielsen, 2015).  Nevertheless, relatively few studies have looked the automation of news distribution and user engagement at the local level. In one exception, scholars looked at the distribution of local news content via Twitter (Meyer & Tang, 2015), but did not focus on the connection between newsroom processes and social media.

In order to better understand algorithmic technology in local newsrooms, data in this study were collected from two primary sources. First, a random sampling of local 663 news websites in 100 communities across the United States was used to create a broad survey of the use of algorithmic technology in local contexts. Copies of the websites of those local newspapers were archived in the summer of 2016 to create a constructed week sample. Qualitative measures were used to assess the degree to which each local news outlet engaged with social media and other algorithmically based distribution channels and to assess the type of coverage, including local news, sports, community events, and politics. Second, and in complement to the overview data, three in-depth case studies were conducted examining specific efforts to develop mobile applications for three separate local news organizations. The mobile application development occurred between 2016 and 2017, and in each case was undertaken by a news organization with fewer than 20 employees. The data for the case studies include in-depth interview data and observations, as well as corresponding metrics from each news organization tracking downloads and usage.

There are three key findings from this study. First, the survey of local news efforts demonstrates that the majority of local news organizations are engaged with algorithmic technology, but show that engagement is generally through third parties as opposed to being native to the organization. Second, the findings of this work demonstrate that the local context exacerbates the ‘black box’ problem of algorithms and news distribution. Based on the case studies, it is clear that third party development of algorithmic technology reduces the role of the editorial process and shifts power away from editors and journalists to programmers. Finally, the aggregate analysis suggests a tension for local news organizations using algorithmic technology; the technology creates a dependence on the very organizations that are fostering increased competition for online audiences. The cannibalistic nature of algorithmic technology at the local level threatens the long-term health of the local news ecosystem. By engaging outside developers, and by relying on third parties to provide the technology necessary to adapt, the local news organizations in this study increase their technological capacity but fail to increase long-term technological knowledge.