Some research from our lab, the Center for Civic Media, because it’s fun and something I’m glad we produced.
In the US, NFL football is more than a sport – it’s a stage on which broader national dramas play out. In the past years, the NFL has brought to national attention conversations about domestic violence, about cheating and fairness and about the ethics of loving a sport that is likely killing its players. With Colin Kapernick’s decision not to stand for the singing of the national anthem during a pre-season football game, starting a wave of similar protests by athletes, a national debate about endemic racism in the US has now become a debate about race, protest, politics and NFL football.
Some years ago, journalist and activist, the late Dori Maynard posed a question to the Media Cloud team: Does sports media use different language to talk about black and white athletes? The question, Dori told us, came from basketball player Isaiah Thomas, who had observed that journalists often described black athletes as physically talented but talked about the intelligence of white athletes. While both descriptions are laudatory, they focus on different aspects of a player’s talents, and enforce long-standing racial stereotypes about intellect and physicality. Could Media Cloud, Dori wondered, put some numbers to these anecdotes?
This isn’t a new research question. Scholars have analyzed the language play-by-play announcers use and have seen the patterns in which white players are praised for intelligence and black players for physical attributes. (See also Rainville and McCormick, 1977 and Rada 1996) Media Cloud gives us the chance to analyze a different corpus, sports stories written after the game, and to examine this possible phenomenon on a different scale. We focused our study on the attention paid to and language used to discuss NFL quarterbacks, the most highly paid and most discussed players on the field.
So do we talk about white quarterbacks as intelligent and black quarterbacks as athletic? Well, like almost everything involving media and race, it’s complicated.
First, we talk a great deal about football in the US media. We analyzed tens of thousands of stories from 478 publications (including US sports websites like NFL.com as well as national and regional sources) over 4 months of NFL regular season coverage in 2015.Despite the prominence of stories like , the vast majority of writing about football discusses this week’s results, next week’s matchup and teams’ strategies for success. As a result, the table of word frequencies when we talk about quarterbacks is heavy on two kinds of words: words that describe gameplay, and words that describe injuries.
We’ve classified each of the 53 quarterbacks who played in NFL games last season as white, black or hispanic (using data from the besttickets unofficial NFL player census, acknowledging that these categories are socially constructed, complex and overlapping.) We then examined what words are associated with coverage of white QBs and QBs of color. In general, white QBs were slightly more associated with action words – ran, threw, leapt – and non-white QBs with words about their health and bodies, their off-field lives and descriptive words, like “dominant†or “judgementâ€. (Our handcoding of the top 250 words associated with QBs, and synonyms for those words, is here.)
We further examined what words were disproportionately associated with white and non-white QBs. For instance, the words “Heisman†and “trophy†were more than three times as likely to appear in stories about black QBs than about white QBs, likely because Heisman winning black QBs Marcus Mariota and Jameis Winston played more last year than white Heisman winner Johnny Manziel. Some of those terms do suggest a focus on the physicality of black QBs:
| Word used more with black QBs | Usage | note |
|---|---|---|
| “Mobile†| 2.48x | |
| “Threat†| 2.46x | (aka: “dual threat†to run or pass) |
| “Legs†| 2.03x | |
| “Runner†| 2.00x | |
| “Scrambling†| 1.97x | |
| “Rushing†| 1.92x | |
| “Sliding†| 1.87x | |
| “Speed†| 1.84x | |
| “Balance†| 1.84x | (may refer to a “balanced offense†as well as to the physical characteristic) |
Words disproportionately associated with white quarterbacks tend to characterize specific scandals and controversies. In most cases, these words describe only one or two quarterbacks, whereas the words disproportionately associated with black QBs often describe multiple players:
| Word used more with white QBs | Note |
|---|---|
| “Deflated†| only associated with Tom Brady |
| “Charter†| only associated with Ryan Mallett missing a charter flight |
| “Court†| |
| “Hormone†| |
| “HGH†| |
| “Jazeera†| An Al Jazeera story about possible use of human growth hormone in the NFL |
Words associated with both white and black quarterbacks, but disproportionately with white QBs also include “domestic†(ie., domestic violence) and partying.
Before concluding that US media is somehow biased against white QBs and their scandals, it’s worth keeping in mind that these terms disproportionately associated with white QBs are highly idiosyncratic – they’re more the portrait of a single player’s struggles than the way a whole group of players are characterized. Moving down in the frequency table to words that appear 1.5x to 5x more with white QBs than black QBs, we find some evidence to support the “white brains, black bodies†hypothesis, but less than we expected.
| Word used more with white QBs | Usage | Note |
|---|---|---|
| “Slipped†| 4.3x | |
| “Slow†| 4.2x | |
| “Prepared†| 2.3x | |
| “Practice†| 2.1x | |
| “Caller†| 1.9x | (“signal callerâ€) |
| “Steady†| 1.7x |
If there’s no racial smoking gun in looking at word frequencies, it may be because, as John Caravalho put it, “No broadcaster or sportswriter this side of Rush Limbaugh is so self-destructive as to blatantly muse on the suitability of a black quarterback.†Reporters may be increasingly sensitive to issues of word choice. But the amount of attention paid to white versus black QBs tells a somewhat different story.
We analyzed how much media attention each of the 53 quarterbacks in our study received. To adjust for the fact that some quarterbacks in our set played very few minutes, we calculated words per minute played, a statistic that ranged from 25.5 words/minute for Titans backup Zack Mettenberger, to 471.4 words/minute for the Cowboys Tony Romo, who suffered a shoulder injury and missed most of the season, to the great dismay of the Dallas press. While Romo is the largest outlier in the set, five other quarterbacks – all white – received unusually high words per minute scores: Brandon Weeden, Johnny Manziel, Landry Jones, Peyton Manning and Tom Brady. The first three – Weeden, Manziel and Jones – played very few games – Jones was a substitute in a single game, while Weeden and Manziel started fewer than 3 games in a 16 game season – skewing these counts. Manning and Brady are “name-brand” quarterbacks, who received additional attention in 2015, Brady for the ongoing “Deflategate” saga and Manning for winning the Super Bowl and retiring.
Comparing a quarterback’s passer rating to his words of coverage suggests that “name brand” quarterbacks are at a distinct advantage in terms of media attention. Six quarterbacks – five white, one black – appear as outliers in this chart. (Romo, who we code as “Hispanic”, didn’t play enough minutes in 2015-16 to have a QB rating.) Peyton Manning, Aaron Rodgers and Tom Brady are all elite quarterbacks who are also recognizable public figures, endorsing products and commanding media attention. (All receive more than $6m in endorsements per year, and rank #1, #4 and #5 in the list of QBs ranked by endorsement money in 2015.) Manziel’s disproportionate attention springs from notoriety – he was benched after videos surfaced of him partying during a bye week – while Andrew Luck had an injury-plagued season that was both poor and widely discussed. The only black quarterback who is an outlier in this set is Marcus Mariota, who outperformed expectations for the Titans, and generated widespread hand-wringing in Tennessee when he was injured late in the season. Notably, the year’s best-rated quarterback – the Seattle Seahawks’ Russell Wilson – is black, and received significantly less attention than worse-rated “brand name” quarterbacks, though average attention for his rating as predicted by our model. Like Manning, Rodgers and Brady, Wilson makes more than $6m a year in endorsements, but his financial success doesn’t lead to disproportionate coverage. Nor does it lead to overcoverage of Drew Brees and Eli Manning, white QBs who were #2 and #3 on the endorsement list in 2015.
Given the messy relationship between performance and attention, we asked whether a naive hypothesis – that sportswriting coverage tracked actual performance – might help answer Dori and Isiah’s question. If black quarterbacks tend to be described as “athletic”, might it be in part because their athleticism is more impressive than that of white quarterbacks?
We looked at two statistics to try to calculate “athleticism”: the 40 yard dash and rushing yards gained by the quarterback. White quarterbacks averaged a little over 4.8 seconds on the 40 yard dash, while black quarterbacks averaged a little below 4.6 seconds. In the NFL, that .25 second gap is an eternity – black quarterbacks, on average, run nearly as fast as receivers, the fastest players on the field, while white quarterbacks are closer to linebackers. That speed apparently matters, as black quarterbacks averaged a little over 200 rushing yards in a season, while white quarterbacks generally had fewer than 50.
This finding about differences in athletic ability by race is obviously heavily loaded, given the long history of racist speech that portrays blacks as fundamentally physically different than whites. We note that the system that results in the presence of more athletic black quarterbacks than white quarterbacks in the NFL is a highly complex one that is deeply embedded in the racial mores of our society. This piece on how modern NFL quarterbacks are made finds that the top 15 quarterback prospects in the 2016 draft overwhelmingly: started playing quarterback by age 9, came from stable families in homes worth at least the median home value, had outside coaching starting in high school, and participated in year round formal 7v7 programs. This kind of intense, adult driven athletic experience is much more common in suburban communities than urban communities. For one example, his piece on the “Hidden Demographics of Youth Sports” lists the five states with the lowest rate of high school sports participation, and four of those five are among the states with the most black households. All of this is to say that this data on the athletic advantage of black over white quarterbacks may or may not say anything about inherent athleticism of black people but almost certainly says something about the deeply racially infused cultural systems that produce modern professional athletes.
Given all of the above, there’s an argument that black quarterbacks are genuinely more athletic – at least in terms of foot speed – than white quarterbacks, and the differences we see in language about quarterbacks may correlate to their performance. That may run counter to suspicions that led Dori to ask her question. But we did find a way in which there’s an apparent racial disparity in coverage: sheer attention.
Only eight quarterbacks broke the 40,000 word barrier in our set, two black, one hispanic, five white. Set the bar at 50,000 and we’re down to four white QBs and Tony Romo. At the highest levels of attention, four “name-brand” quarterbacks (Rodgers, Brady, Manning, Romo) and one screw-up (Manziel) dominate discussion of football in 2015-6. Elite black QBs – Russell Wilson, Marcus Mariotta, Cam Newton – received more attention than mediocre quarterbacks, but less than name brand, endorsement laden white QBs, despite in Wilson’s case, significantly superior performance.
Is there a racial bias in sportswriting about the NFL? Probably.That bias may be related to which NFL players gain endorsement contracts and widespread celebrity, and which ones fall short of expectations to reach that elite level. It’s difficult to entangle causality, though – all but one of these “name brand†QBs are white, and we may pay attention to them because of their celebrity, which correlates only partially to their superior athletic performance, and may correlate more closely to their race.
We will be updating our study at the close of the 2016-7 NFL season, and are looking forward to seeing whether Kapernick’s protest challenged the attention patterns we saw in the previous season.
This post was written by Ethan Zuckerman in collaboration with Allan Ko, Rahul Bhargava, and Hal Roberts. Allan Ko produced the graphics and conducted the quantitative research.