How The Post gathered and analyzed data for its series on Black NFL coaches

As the National Football League faces litigation alleging racial discrimination at its highest coaching levels, reporters at The Washington Post wanted to know: Does the experience of Black head coaches differ from that of their peers?

Over seven months, Post reporters built data sets with the goal of examining this question. The Post also contacted, either directly or through a representative, every living Black current or former head coach. Sixteen of the 24 agreed to be interviewed, including 14 of the 19 who held full-time positions. (Dennis Green, the second Black head coach in the modern era of the NFL, died in 2016.) Post reporters traveled the country, visiting homes, offices and team facilities to conduct in-person interviews with those 16 and also interviewed dozens of their colleagues.

The Post asked the NFL multiple times to provide data to aid reporters’ efforts. A spokeswoman said in July that the league could not provide any of the data The Post requested, which included aggregate demographics on a season-by-season basis for head coaches, coordinators, position coaches and players.

The Post also asked the NFL to verify coaches’ dates of birth and racial identification. In the days before this story published, the league provided The Post with four tables that showed the aggregate demographics of head coaches, offensive coordinators, defensive coordinators and special teams coordinators over a 21-year span — not separated by season.

Without data directly from the NFL, The Post gathered information from multiple sources for the league’s 191 head coaches since 1990, the year after Art Shell became the first Black coach in the modern history of the league. The Post compiled their race, date of birth and information about their coaching careers, such as how many seasons they spent as a head coach and their career record, from sources that include statistics website Sports-Reference.com, team bios, public records and the coaches themselves.

Reporters also collected coaches’ year-by-year results and full job histories, allowing The Post to analyze each coach’s network, identifying which coaches each had worked under and those who later worked under him.

Identifying coaches’ races

Reporters first needed to identify the race of each coach. To do this, The Post relied on numerous sources: the NFL’s diversity and inclusion reports from 2013 to 2022, which include the racial identification of select coaches; a data set from researchers at Arizona State University’s Global Sport Institute; and information provided by University of Pennsylvania professor Janice Madden, who, with Matthew Ruther, has studied the NFL’s Rooney Rule.

Each data set used slightly different methods to identify race. The lead researcher for the NFL’s diversity and inclusion reports, C. Keith Harrison, said his effort used “multiple credible sources” to pinpoint racial identities, but he wouldn’t provide additional details, and an NFL spokeswoman did not respond to questions. The Global Sport Institute’s racial identification relies on publicly accessible sources.

Madden used some visual evaluation to determine race, a practice she said mirrored “the evaluations of the hiring/firing parties whose behavior and choices we were studying.” Only 25 coaches in The Post’s data set, all of whom worked in the 1990s, were solely identified through Madden’s visual approach. Madden identified all of them as White.

For most analyses, The Post grouped coaches into three categories: White, Black and other coaches of color. The coaches in the other category are Mike McDaniel (biracial), Robert Saleh (Lebanese American), Ron Rivera (Latino) and Tom Flores (Latino).

McDaniel has a Black father and a White mother. According to a Miami Dolphins spokeswoman, he identifies as biracial. The NFL diversity and inclusion report lists him as multiracial and league officials did not include him in their count of Black coaches. Through the team spokeswoman, he did not respond to questions about how he would prefer to be categorized by The Post.

Saleh, of the New York Jets, told The Post he identifies as a minority. According to the U.S. Census, people of Middle Eastern descent are considered White, but Saleh is described as a coach of color in the NFL’s reports, so The Post categorized him with other non-Black coaches of color.

Where possible, The Post used coaches’ racial self-identification. For instance, Herm Edwards is biracial but has identified himself as Black, and Brian Flores, who is listed in the NFL’s reports as Honduran American and Afro Latino, also identifies as Black. The Post identified both coaches as Black in its analysis.

The Post relied on data from the Institute for Diversity and Ethics in Sport (TIDES) at the University of Central Florida for the demographics of NFL players by year. In some years, the league did not provide data to TIDES. The Post worked with data from the Global Sport Institute to determine the demographics of coordinators.

Coaches’ career paths

The Post pulled each coach’s job history from Sports-Reference, then verified each position through public sources, adding and adjusting numerous roles. Job titles were grouped into general categories such as coordinator or low-level assistant for analysis purposes. NFL playing time was counted only in years when a player appeared in a game.

Post reporters categorized a primary coaching area — offense, defense or special teams — based on a coach’s career path. In general, if a coach spent at least three years longer in roles of a certain area than all others, he was designated with that coaching area. All data related to performance as a head coach came from Sports-Reference. All data related to win totals and winning percentage is for the regular season unless otherwise noted.

In general, the average number of years spent in a role represents the average of all coaches, even if some did not hold that role. For instance, only 28 percent of first-time NFL head coaches had previously served as college head coaches, and they held their roles for an average of seven years. But The Post’s analysis includes those who never held college head coaching roles, bringing the average down to about two years.

The Post relied on news reports to determine coaches’ postseason outcomes and whether departures should be classified as “fired” or “otherwise left.” Coaches who retired or resigned, even amid pressure or turmoil, are included in the “otherwise left” category.

Coaches’ ages for a given season are determined based on their age as of Jan. 1 of that year.

Network analysis

The Post used network analysis, a technique that represents relationships between different members of a group, to analyze professional connections between NFL coaches. Limiting the analysis to full-time coaches who coached games in or after the 1990 season, including the 32 coaches coming into the 2022 season, The Post visually and quantitatively represented relationships between these coaches.

The Post limited its analysis to relationships at the NFL level in which one person served as a full-time head coach and the other was employed by him as a coach or adviser. For example, in 1996, Lovie Smith was the linebackers coach for the Tampa Bay Buccaneers, when Tony Dungy was their head coach, so Dungy and Smith are connected in the network. Smith worked under Dungy for five years, so their connection has a weight of five.

Interim coaching analysis

Coaches who served in an interim role but were not credited with results are not included in The Post’s analysis.

Gary Moeller was elevated to the Detroit Lions’ head coaching job during the 2000 season following the resignation of Bobby Ross, but he is not included in the analysis because he signed a multiyear contract.

Aaron Kromer and Joe Vitt were interim coaches for the New Orleans Saints in 2012 while Sean Payton served a one-year suspension. They are included in The Post’s analysis related to the demographics and overall performance of interim coaches, but they are not included in the analysis or charts related to whether interim coaches were hired for the full-time job. Payton returned the following season, so Kromer and Vitt did not have an opportunity to be promoted.

Vitt also served as the interim coach of the St. Louis Rams in 2005 when Mike Martz missed the final 11 games because of a health issue. Martz was fired after the season, so The Post included Vitt’s interim tenure in its analysis because his stint resembled the audition period that most interim coaches experience.

When comparing performance of interim coaches, The Post used their combined record to calculate a winning percentage rather than equally weighting each coach’s winning percentage in an average regardless of how many games he coached.