NPR's On-Air Source Diversity: Some Improvement, More Work To Be Done : NPR Public Editor More of NPR's sources are black, but female, Latino, and Asian sources show little to no gain over three years.

NPR's On-Air Source Diversity: Some Improvement, More Work To Be Done

A 2015 study of the racial/ethnic diversity of sources on NPR's weekday newsmagazines shows this breakdown. Annie Johnson/NPR hide caption

toggle caption
Annie Johnson/NPR

A 2015 study of the racial/ethnic diversity of sources on NPR's weekday newsmagazines shows this breakdown.

Annie Johnson/NPR

Results are in from the third year of NPR's sourcing project—designed to understand and ultimately improve the gender, geographic, and racial and ethnic diversity of people heard on NPR as outside sources of news and opinion. The news is mixed.

In fiscal year 2015, which ended Sept. 30, there was a notable increase, compared to two years earlier, in the percentage of black sources, and an incremental, statistically insignificant increase in the share of female sources. Most disappointingly, there was virtually no change in the share of Latino sources.

Asian sources improved slightly (again, a statistically insignificant change). The third-year results reflect a variety of experiments to improve source diversity that took place in 14 newsroom divisions over nine months.

The impressive and challenging source diversity undertaking, which began in the fall of 2012, is led by Keith Woods, NPR's vice president for diversity in news and operations. It is a major part of how NPR hopes to reach its oft-stated goal to "sound like America," which is related to its goal of broadening the diversity of those who listen. Currently, 8 percent of NPR listeners are Latino, 7 percent are black and the vast majority of the rest are white. I wrote about some of the findings from the first two years of the sourcing project in a previous column.

I asked for an advance look at the numbers, which are being released this week to the NPR newsroom and member stations. My office gets many questions from listeners about the voices and perspectives they hear on the air—or in some cases, don't hear. Numbers can't tell the whole story, but they can provide some insight; after all, no listener can hear everything NPR puts on the air.

First off, a caveat: The project was limited to an analysis of the sources heard on NPR's two largest weekday newsmagazines, Morning Edition and All Things Considered. The results might look a little different if they also had included the weekend newsmagazines, or the online-only reporting found at, or the other programs NPR produces or distributes (for example Here and Now, and Latino USA.) Moreover, the project is based on a sample of four randomly selected composite weeks of data (totaling 769 stories and 2,080 sources this year; the three-year total is roughly 2,400 stories and 6,200 sources). A full year of data might also produce different results.

The sources were characterized using publicly available data, reporters' notes or in some cases by calling them after the broadcast; no guesses were made by, say, extrapolating from a source's surname. A small number of voices that were not characterizable were excluded.

The findings are extensive, including a breakout of international versus domestic sources and a separate look at the people who were asked to appear on air as subject matter experts (as opposed to all sources, including the general public). But I am focusing here on the overall numbers, which are illuminating.

Looking at race and ethnicity, the findings showed a decline in the overall percentage of white sources, to 73 percent, in fiscal year 2015, compared to 80 percent in fiscal year 2013. As the dominance of white voices dropped, the share of black sources rose to 11 percent, from 5 percent, and the share of Asian sources rose to 8 percent, from 5 percent. But the share of Latino sources stayed flat, at 6 percent each year.

Diversity In Sourcing By The Numbers

The sourcing project examined the diversity of sources on NPR's weekday newsmagazines. All charts except the first show results for FY 2015.

Gender, too, was effectively constant, with male sources outnumbering female sources two-to-one. Overall, in fiscal year 2015, 30 percent of NPR's sources were female. (Transgender and multi-gender sources were too small of a percentage to show up in the findings.)

Woods said he was "generally pleased with the direction that this is going," noting the increases in the share of black on-air sources, as well as the percentage of "subject matter experts" who are people of color. He said he had "hoped for better news on our coverage of women, on our inclusion of women."

His biggest concern, he said, is that NPR has "not substantively changed things with our coverage of Latinos. Flat is not the new up. Flat is the new down when you consider what's happening with the population in the country." 

In terms of geographic diversity, year three results included far fewer sources from Washington, D.C. That may have been partly due to the stories that were prominent during the study period, although newsroom managers in recent years have also been pressing journalists to lessen their reliance on "beltway" sources from the government and policy institutes around town, which are heavily populated by white men. The traditional heavy reliance on government, policy and business and economics stories—all fields still dominated by white men—has also created some of the challenge NPR faces, Woods said, as has NPR's longtime fondness for interviewing professors and other journalists.

Woods said he believes the increase in black sources is "partly a response by journalists to the emphasis on greater diversity." Another contributing factor, he said, was likely "the number of national stories over the last fiscal year that were about racial conflict, especially the major stories about police shootings."

"The good news for NPR is that we're covering something that is important to people who are black primarily, and Americans more broadly," Woods said. "That's a big step in diversity; that you cover what's important to the people that you wish to have in your audience. But also important is, if you look at those expert numbers where we went from 16 percent people of color as experts to 25 percent in those three years, what you see is that from these stories that are impacting people of color, that NPR is going to the experts of color to talk."

Why were the numbers on Latino sources so low, and unchanged?

"It underscores we will have to take some direct steps ourselves" to cover Latinos, Woods said, as opposed to relying primarily on breaking news stories to bring in Latino voices. NPR could also increase its coverage of religion and education topics, he said, where Latino sources are better represented than in NPR's overall news coverage. Or, Woods said, NPR could shift some government coverage away from what he called "incremental politicking" and instead spend more time talking to people in localities affected by those political decisions—a strategy I would endorse for more reasons than just improving the diversity of sources.

There is much more to glean from this study, and in coming weeks executives will be pulling the numbers apart and creating strategies to move forward, Emma Carrasco, NPR's chief marketing officer and senior vice president for audience development, told me. "For me it's always about where our gaps and opportunities are," she said, and having the data makes it possible to understand that more clearly.

Improving the number of Latino sources is a "huge opportunity," she said, adding that while the data is most importantly applicable to the newsmagazine editorial teams, "there is a lot here that all of us working on various aspects of content need to pay attention to."

Michael Oreskes, NPR's senior vice president of news, in an email called the data "very encouraging. It shows that we can make change. We reduced the number of our sources who are inside the beltway and increased the number who are African American. We still have a lot of work to do to get where we need to be to really capture the extraordinary diversity of this country. But these experiments have shown us techniques that work and we can build on that."

Overall, Woods said, given that NPR was able to make strides over the three years of the project, "I can't be anything but profoundly optimistic about the trend in our numbers."

And here's an important note about the newsroom's optimism: NPR executives have smartly not set a numerical goal that they hope to achieve in terms of source diversity, which Woods said would create a "sense of artificial imperative." Instead, the company emphasizes the need to create a newsroom culture where the sources of news and opinion are more in line with the makeup of America.

As for me, I am impressed that NPR undertakes such a serious self-examination, and I am eager to hear how newsroom executives will use the findings to continue improving the diversity of voices on NPR shows. Knowing the numbers is a good start, but it's just that. I have already written twice about what I think is an urgent need for NPR to diversify its regular commentators, who are overwhelmingly white men.

In an upcoming post I will look in more detail at some of the ways the newsroom is changing its approach to sourcing, as well as some of the challenges. I will also update the staff diversity numbers that my predecessor published in June 2014.

Part II Dec. 15, 2015

This post was updated to include a link to the second part of this analysis, which can be found here.