In January 2021, the Journalism and Media Ethics Program at the Markkula Center is set launch a new and timely prototype project with funding from the Google News Initiative (GNI). GNI announced details of the program here.
Particularly in 2020, newsrooms across the country have made impassioned declarations about their commitments to diversity, equity, and inclusion (DEI). But local newsrooms are increasingly resource-constrained, which makes manual diversity audits of sourcing and quoting practices impractical. Human audits are valuable, but time-consuming, expensive, and retrospective, usually done annually. Human audits also do not produce an everyday “nudge” factor (a concept from behavioral economics) to encourage behavior change.
The Journalism and Media Ethics program’s solution is to offer real-time/on-demand visualizations for source-diversity proportions in quotes through an automated dashboard. We plan to release this as a prototype WordPress plugin and a web monitor application. One critical value add is that reporters can use it on-demand as they draft their stories to ensure they are equitably quoting sources. It will also provide source-diversity analytics for published work around the taxonomy a newsroom’s content management system is already using. The web application will offer the same visualizations at the site level.
Subbu Vincent and Anita Varma, program leaders for Journalism and Media Ethics at the Ethics Center, will work with SCU engineering students to build out the prototype and training material. Associate Professor of Computer Science and Engineering, Yi Fang, will lead the work to prototype community representation detection around larger demographics, such as BIPOC versus non-BIPOC. The analytical framework used to build the toolkit is based on a synthesis of journalism ethics vocabularies and design interviews with local Bay Area news reporters.
The plugin will be released for testing and usage to a small group of early adopter newsrooms (to be announced in early 2021) starting summer ‘21.
"We're excited to support journalists' important work with an automated tool for reporters to visualize the representativeness of their stories--prior to publication--with respect to who they are and aren't sourcing. We strongly believe that journalism ethics tools need to account for current realities of limited resources that can make it hard for journalists to 'do it all,’” said Anita Varma, assistant director of Journalism and Media Ethics at the Markkula Center for Applied Ethics.
“Let’s say a reporter has 800 words drafted for an enterprise story, but it’s not done yet. From her draft, she can quickly pull up the list of people quoted, their quotes, the gender proportion, a community representation proportion, expert quotes vs. non-expert quotes proportion, etc. She can see how often she is quoting a particular source. She can see her source diversity proportions across all recently published stories. The source-diversity dashboard tool will show all of these numbers on-demand,” said Subbu Vincent, director of Journalism and Media Ethics, outlining a use case for the tech and methods toolkit.
The system will use Natural Language Processing (leveraging the Stanford CoreNLP system, adapted and customized for journalistic writing) to identify quotes, providers of quotes, their gender, titles, and organizations. The longer-term development plan is to add community representation data to quote visualizations in the second version by late 2021.
The program also acknowledges the City University of New York’s NewsQ Initiative for their 2020 grant to the Markkula Center for Applied Ethics for our Detecting the Boundaries of Journalistic Behavior Detection Online project. This helped us build the back-end NLP infrastructure software for quote audits that we are leveraging for the DEI toolkit prototype.