How to analyze for inclusion: Our experience with KPCC/LAist

Many news organizations are doing source tracking, and knowing the percentage of women represented in a content is undoubtedly valuable. But it’s even more helpful to know the proportion of women cited in politics coverage as opposed to education because it can lead to more targeted actions.

Since 2019 Impact Architects has partnered with nearly a dozen publications — nonprofit, public media, and commercial — to conduct source and inclusive content analyses to go a step further and generate insights to guide publications to be more inclusive in their sourcing and content creation. Our analyses have focused on sources as the main unit of analysis and the point of possible intervention for publications to make a change, but go beyond tracking sources to identify interesting trends and opportunities by breaking down source type (e.g., gender, race and ethnicity, or profession) by story characteristics, like story type (explanatory, investigative, etc.) and topic (politics and policy, health, education, etc.).

In mid-2022, we were excited to have the opportunity to work with KPCC/LAist (a part of Southern California Public Radio) to take the next step in developing an inclusive content analysis that focused specifically on content. KPCC/LAist had specific questions we hoped to be able to answer:

• How are KPCC/LAist centering (or not centering) people in story-telling? Is content from a resident’s point of view or an institutional point of view?
• Is KPCC/LAist’s use of images inclusive and relevant to the story?
• How often does KPCC/LAist frame a story or issue as a problem versus a solution?
• How inclusive is KPCC/LAist’s language?

Through collaboration with KPCC/LAist, we developed a methodology to answer these questions using a randomly selected, representative sample for four content areas: the AirTalk talkshow, AM and PM newscasts, the Morning Brief newsletter, and digital publications from the LAist.com published from July 1-December 31, 2021.

How did we measure the inclusivity of content?

First, we broke down the questions into three big themes for coding: Framing, imagery, and inclusive language. The questions about framing and imagery are general and can be taken and applied for just about any news organization, as they emphasize headlines and lead images. The questions about inclusive language are more tailored to KPCC/LAist, as we relied on their Dialogue style guide, which is a codified and living reference for KPCC/LAist journalists. You can read the full, detailed results here.

What we found

Through all of our inclusive content analyses, our goal is to provide useful information to our clients, not to tell them whether or not they’re “doing well” or “doing poorly,” even though they might draw those conclusions themselves. Interpreting the results is usually based on where the organization and/or team is and where they want to be.

Framing

Regarding framing, we found very few examples of hyperbolic headlines for KPCC/LAist. However, we did find headlines that didn’t directly relate to the theme of the story. Most often, these were in the Morning Brief newsletter, which exclusively used a headline structure of: “Lead story, select bullet point, leisure activity.” As a result, the headlines tended to be non-descriptive, even if they passed the threshold of “relating to the topic,” while at other times, the structure that puts three unrelated things together could end up making light of serious topics, as in “Guns, Abortion, and Tacos” or “LA Hate Crimes, An Apple 1, And Pizza.” The structure might have been an attempt to optimize search. In an unsystematic look at Morning Brief headlines published after the conclusion of our audit, it seems that the practice has changed.

Another component of framing is people-centered language. We found that 65.5% of stories analyzed centered the voices of people rather than institutions (which we defined as sources that appear to speak on behalf of an institution). KPCC/LAist can determine whether or not this is where they want to be in this respect, but they can also use the finding to ask why and when institutional voices do appear, as the assumption is not that they should never be represented but that they shouldn’t be the predominant voice.

Imagery

We found more images that did not directly relate to the theme of the story than we did headlines. Some of the images we flagged as not relating to the story seemed to be the result of using shortcuts that leave out important thematic context. For instance, an image of a tent encampment for a story about a nonprofit that serves people experiencing homelessnes, but not about the people themselves, or a story about a support center for the trans population that features a pride flag. The last example would be like a story about a veteran’s center that features a POW/MIA flag.

Making changes in images is more difficult because it requires greater capacity and resources. Headline writing practices and story framing can be changed with training and establishing specific standards and a mindset in the newsroom, but KPCC/LAist might have limited options for including an image in a news story, especially brief ones that don’t require a lot of reporting. Because of capacity and available resources, stock and loosely related images might persist, but the insight should at least lead editors, producers, and journalists to think more deeply about image selection.

Inclusive language

In our entire analysis, we flagged at least one example of non-inclusive language for every category in the rubric, but our research didn’t find any trends regarding non-inclusive language. Instead, these instances tended to be one-offs. Still, KPCC/LAist can ask “why” and “how” questions for each example: Why did we say it this way? How can we say it differently?”

For example, when a newscast said “Blacks comprise 90% of Rodeo Drive arrests,” the answer to the “why” question might be that it’s fewer words and every fraction of a second counts in a newscast, while the answer to the “how” question might be to say “Black people” and find a fraction of a second to cut elsewhere.

What you can do

We at Impact Architects were gratified to work with KPCC/LAist to expand our content auditing toolset and develop a fresh methodology. As noted earlier, the questions regarding framing and imagery can be applied to most news organizations. When answering “how inclusive is our language?” for KPCC/LAist, we really were answering “are we living up to our own standards?” Different organizations might have different approaches, so these questions could be tailored to fit different standards.

For any organization that wants to think harder about framing, imager, and language, there are resources available. KPCC/LAist’s Dialogue style guide is indeed one of them, Vox Media launched Language, Please, a free resource for journalists, and Resolve Philly’s Reframe includes resources for word choice and framing (and has an excellent newsletter).

And, of course, we’d love to work with more organizations to generate insights and useful information about news content. If you’re interested in partnering with us for this type of work at your organization, send me an email: eric@theimpactarchitects.com.