Massive datasets of communication are challenging traditional, human-driven approaches to content analysis. Computational methods present enticing solutions to these problems but in many cases are insufficient on their own. We argue that an approach blending computational and manual methods throughout the content analysis process may yield more fruitful results, and draw on a case study of news sourcing on Twitter to illustrate this hybrid approach in action. Careful combinations of computational and manual techniques can preserve the strengths of traditional content analysis, with its systematic rigor and contextual sensitivity, while also maximizing the large-scale capacity of Big Data and the algorithmic accuracy of computational methods.
19 March 2014 in Journal of Computer-Mediated Communication
Amid growing threats to journalists around the world, this study examines the nature of online harassment, the types of journalists most likely to experience it, and the most common forms of response to such abuse. Through a representative survey of U.S. journalists, we find that nearly all journalists experience at least some online harassment but that such harassment is generally infrequent overall and especially in its most severe forms. Nevertheless, online harassment against journalists disproportionately affects women (particularly young women) and those who are more personally visible in the news but not necessarily those who work for larger newsrooms. Moreover, it is clear that the more often a journalist is harassed online, the more likely they are to take a dim view of the audience by seeing them as irrational and unlike themselves, and to perceive interaction with them as less valuable. Additionally, as greater targets of the worst forms of abuse, women face a greater burden in deciding if and how to respond to online harassment. Conceptually, this article advances the literature on journalists and audiences by extending the concept of reciprocal journalism, which emphasizes individual-level perceptions that shape the quality of person-to-person exchanges. We explore how the experience of online harassment may complicate the way that journalists think about and act toward their audiences, offering a window into the downsides of encountering audiences online.
A number of social, technological, and economic shifts over the past two decades have led to the proliferation of audience analytics and metrics in journalism. This article contends that we are witnessing a third wave toward the rationalization of audience understanding and distinguishes between audience analytics (systems that capture information) and audience metrics (quantified measures output by those systems). The body of literature on analytics and metrics in the context of news production is then synthesized across the ABCDE of news production: attitudes, behaviors, content, discourse, and ethics. That synthesis leads to an overarching conclusion that while contemporary journalism is not being driven by quantified audiences, both audiences and quantification are playing far more prominent roles in news production than in the past. Scholars and practitioners have also become less pessimistic about analytics and metrics over time, recognizing more nuanced effects and prosocial possibilities. Finally, important gaps in the literature are identified and new research directions proposed to help address them.
This article examines the features of day-to-day data journalism produced by The New York Times and The Washington Post in the first half of 2017. The content analysis evaluates story characteristics linked to the concepts of transparency, interactivity, diversity, and information provenance. It finds that the data journalism produced by those outlets comes from small teams, focuses on “hard news,” provides fairly uncomplex data visualizations with low levels of interactivity, relies primarily on institutional sources and offers little original data collection, and incorporates just two data sources on average in a generally opaque manner. This leads to the conclusion that “general data journalism” still has a long way to go before it can live up to the optimism and idealization that characterizes much of the data turn in journalism. Instead, contemporary day-to-day data journalism is perhaps better characterized as evolutionary rather than revolutionary, with its celebrated potential to serve as a leap forward for journalism and engender greater trust in it remaining untapped.
9 April 2015 in The Annals of the American Academy of Political and Social Science
To deal with ever-larger datasets, media scholars are increasingly using computational analytic methods. This article focuses on how the traditional (manual) approach to conducting a content analysis—a primary method in the study of media messages—is being reconfigured, assesses what is gained and lost in turning to computational solutions, and builds on a “hybrid” approach to content analysis. We argue that computational methods are most fruitful when variables are readily identifiable in texts and when source material is easily parsed. Manual methods, though, are most appropriate for complex variables and when source material is not well digitized. These modes can be effectively combined throughout the process of content analysis to facilitate expansive and powerful analyses that are reliable and meaningful.