16

The Black-White Thing: Racial Biases in NBA Scouting Comparisons

The following is a guest post by Will Garcia. Here’s a little about Will:

I’m just a dude who wishes he could dunk. Instead, I can kinda work an Excel spreadsheet. Follow me @WillTheHoopsGuy or shoot me an e-mail at willthehoopsguy (at) gmail.com to talk shop!

Enjoy! -Ed.

Mandatory Credit: Kelley L Cox-USA TODAY Sports

“Knows [his] limitations and doesn’t make a ton of mistakes … Plays with a lot of energy … Good motor … Has interesting role-playing potential.”

via Miles Plumlee | NBADraft.net

If this description of Miles Plumlee somehow managed to fit in “scrappy,” “gym rat,” “team defense,” and “offensive awareness,” it may have been the whitest description of a basketball player to ever grace a scouting report.

This isn’t a knock on Tyler Ingle, who wrote the report in 2012. He definitely did his job well in describing Plumlee, who made the World Cup roster this year had a breakout season last year in Phoenix largely because he knows his strengths and weaknesses, doesn’t make many mistakes, has a good motor, and is a solid role player.*

This is more of a knock on myself, who read that and thought, “So…he’s a white guy.”

Racially-coded language has long been a problem in American professional sports, as Alex Diamond’s scarily-good-for-a-Bachelor’s-thesis “The Construction of Race in Professional Basketball” points out. The framing of white athletes as ‘skilled and fundamentally sound, despite a lack of physical tools’ and black athletes as ‘athletically superior, but lacking in intangibles and skills’ began as a coping mechanism for racist white sports fans to be able to reconcile the performance of black athletes in the early 1900s (p. 11-12).

These stereotypes came to a head in the early 1980s, with Larry Bird, Kevin McHale and the predominately white Boston Celtics playing a ‘smart, fundamentally sound, highly-skilled’ brand of basketball against Magic Johnson, Kareem Abdul-Jabbar and the flashy, fast-paced, and predominately black “Showtime” Los Angeles Lakers. As Diamond argues, the way that these teams and players were viewed fit hand-in-hand with the time period’s views of race (p. 23).

Of course, these constructions were largely BS.

The “Showtime” Lakers won with skill and by being a team of basketball braniacs. Pat Riley is an established genius of a coach, Michael Cooper and Byron Scott both wound up securing head coaching gigs, and Kareem is the go-to guy that today’s big men (like Dwight, Joakim Noah, and Roy Hibbert) see to improve their games.

As for Magic, the guy literally wrote the book released the DVD series on fundamentals.

Photo courtesy of Brand A Productions and Fundamentals of Basketball, LLC

Photo courtesy of Brand A Productions and Fundamentals of Basketball, LLC

In Boston, the Celtics were arguably playing faster than their West Coast counterparts, shooting more field goals and free throws than the Lakers during the 1984-85 season.

The “white guy, smart guy, black guy, fast guy” myth was so pervasive that it was able to be reinforced using teams and players who proved it wrong. And it hasn’t gone anywhere; even today, players are viewed though a racial lens before they step foot on an NBA court.

Photos courtesy of Al Tielemans (left) and Lane Stewart (right) of Sports Illustrated

Photos courtesy of Al Tielemans (left) and Lane Stewart (right) of Sports Illustrated

Ethnic distribution of player comparisons

I wanted to take an objective look at how race and racial biases affect the perception of NBA prospects. To do this, I began by looking at player comparisons from 565 scouting reports of prospects drafted between 2005 and 2011 from two major scouting websites. I took note of the race/ethnicity, continent, and career average win shares of both the prospects and the players they were compared to.

Unsurprisingly, white guys (n=60) were compared to white guys:

race4

White European guys (n=43) were compared to white European guys:

race5

Alarmingly, black American NBA draft prospects (n=397) were compared to other black American players at a 96% clip!

race6

These findings show that player comparisons are drawn along racial lines for all groups of prospects. Indeed, the “eye-test” unfortunately takes into account the way that players look.

But are the comparisons valid?

The relationship between comparison accuracy and race

To see if the accuracy of player comparisons was skewed along racial lines, Ian Levy from Nylon Calculus (formerly Hickory-High) graciously let me use his Similarity Scores, an awesome database that compares the college production of draft prospects from 2010-2014 to the college production of NBA players and assigns a score based on how similarly they performed. For a rough guideline, 1000 is the maximum score (literally identical production, age, and size), 800 or so is meh, and 667 is Kelly Olynyk-John Wall.

race7

Out of 98 player comparisons for black NCAA players and 10 comparisons for white NCAA players, the similarity scores were similarly OK. Keeping in mind the small sample size of white prospect comparisons, it’s difficult to say if the slight advantage of white player comparisons is indicative of any long-term trends, but it’s encouraging that the two aren’t that far apart and that both numbers indicate that the player comparisons aren’t entirely baseless.

race8

Similarly, there was almost no disparity between how much black American (n=377) and white American (n=55) prospects were outperformed by their comparisons according to Basketball-Reference’s version of the popular win shares metric. Both groups averaged about 1.7 fewer win shares per season than their supposed basketball doppelgangers, with white European prospects (n=33) clocking in slightly worse at about 2.5.

(Funnily enough, when I removed instances of European big men getting compared to Dirk, they fell right in line with everybody else at 1.7. Scouts, cut the Dirk comps. Please.)

Up to this point, we’ve established that race definitely plays a role in how players are perceived and that player comparisons are somewhat accurate, if overzealous, ways to see how a player may perform in the NBA.

But we haven’t answered the big question.

Is there merit to basketball’s racially-charged stereotypes?

To tackle this question, we first need to look at the relationship between race, height, and sports.

Screen Shot 2014-08-29 at 10.57.51 AM

An area of note that doesn’t get brought up in conversations of race and basketball is the fact that white players are disproportionately power forwards and centers, whereas black players are underrepresented at center and more likely to be guards and wings.

This may be due to a funneling effect that pushes average-sized white athletes away from basketball. As this Pew study indicates, two-thirds of black Americans name either football or basketball as their favorite sport. For white citizens, however, getting to two-thirds takes football, basketball, baseball, and NASCAR, and that’s not even counting other white-dominated sports like hockey, golf, and tennis.

A 6’3”, 200 lb white athlete is more likely to be on a baseball field, auto track, tennis court, hockey rink, or golf course than his black peer, and because of the way these sports are played, being average-sized is a good thing.

But if you’re 6’11”, it doesn’t matter what color you are—your abnormally large butt is getting on that basketball court whether you like it or not. Athletically, it’s about all you’re good for.

And it’s a huge advantage; as Bill Marsh notes in the New York Times, one in seven American 7-footers play in the NBA. With odds like that and a lack of other options that make sense for their body type, it’s not surprising that white big men are blocking dunks and layups instead of defensive lines or slapshots.

The positional disparities between white and black players are important to take note of because they impart statistical differences for the groups as a whole. I was a bit shocked when I saw how much white players were outrebounding black players on average, and vice versa with assist numbers.

Of course, once we control for position (in this case, PFs and Cs are the only positions with a large enough sample size both ways to draw any sort of conclusions from), we find that the players score very similarly, albeit with differences in midrange (16 ft. to 3pt. line) and free throw shooting.

race11

The shooting differences may be explained by the relatively small (n=35) sample of white bigs having more than its fair share of knockdown shooters. With a sample size that can be easily skewed by three or four more great shooters than average, having Matt Bonner, Steve Novak, Ryan Anderson, Luke Babbit, Josh McRoberts, Matt Bonner, Kelly Olynyk, Spencer Hawes, Jon Leuer, Ryan Kelly, and Kevin Love in the sample obviously played a big role in the differences we see. Out of a mere 35 players, it’s not that unlikely that there are three or four more elite shooters in this group than there probably “should” be.

Alternatively, the disparity in shooting could be seen as a stereotype threat against black players. A stereotype threat is a situation that puts people in a situation where, if they perform poorly, they confirm harmful stereotypes about their ethnicity, sex, gender, sexual orientation, or other demographic group. Unsurprisingly, it carries over to athletics. As a study from Princeton and the University of Arizona shows, white students performed worse at golf when it was framed as “a diagnostic of ‘natural athletic ability’” whereas black students underperformed when it was called a “diagnostic of ‘sports intelligence.’”

If shooting is framed as a thing that “white guys are better at” (which, let’s not kid ourselves, it totally is), then black players face a stereotype threat when they shoot perimeter jump shots. This could lead to a decrease in performance.

One other interesting thing of note is that black and white big men rebounded virtually identically, proving the “good-motored white guy” stereotype as a myth.

Speaking of myths, arguably the biggest racial myth in basketball is that “white men can’t jump”.

It’s just as false as every other racial myth we’ve encountered thus far.

race12

Using NBA Draft Combine data from Draft-Express, we see that black (n=211) and white (n=45) big men performed almost identically in every drill.

Concluding thoughts

Race and basketball have a relationship that is founded on falsehoods and rooted in a time period defined by bigotry, but still partially reflects the way we feel about blackness, whiteness, and the role masculinity plays between both groups. Like the very concept of race itself, the pillars supporting the notion that basketball is a game inherently divided by race are demonstrably false.

Unfortunately, the attitudes and outlook reinforced by this “racial common sense” don’t rely on truth for their continued existence. Instead, they survive by going unchallenged and being passively accepted.

Our ability as a game (and as a society) to give people the respect they deserve for their talents, efforts, and hard work depends on this changing.

*An earlier version of this piece mistook Mason Plumlee for his brother Miles and resulted in a mix-up about Team USA roster affiliation. Whoopsie! All better now!

Hardwood Paroxysm

  • NotoriousCPC

    Final chart says it all.

  • Bob

    What about guards and wings? What about wingspan and standing reach? This is nice but not a finished product.

    • http://www.hoopsthink.com James Kerti

      “PFs and Cs are the only positions with a large enough sample size both ways to draw any sort of conclusions from”

    • Will Garcia

      There were only 11 guards and about 20 wings. Not nearly enough to produce anything definitive.

      Wignspan and standing reach are important, but the stereotype exists predominately in relation to athleticism, not body dimensions.

      It’s still definitely something worth looking into, though.

      • Bob

        Wingspan is directly related to athleticism, which is where you unfortunately missed the mark. Wingspan probably matters more than anything listed on here. This article has a great message, but I do think some African Americans are more athletic based simply on their natural bodies and there is absolutely nothing wrong with that. There are biases, and some people just have more natural skill.

        • Will Garcia

          Just ran the numbers on standing reach, and I got 108 inches for black American big men prospects and 107.5 for white American bigs.

          Would that be considered statistically significant? Probably.

          Is it actually significant to actual basketball? No.

          • Bob

            I don’t care about bigs. I want guards and wings regardless of sample size. I’ll tell you what: there’s a reason why the “white” sample size is so small…

          • Will Garcia

            ” I want guards and wings regardless of sample size.

            If you want that, you’re more than welcome to go to the Draft Express database for combine results, found at http://www.draftexpress(dot)com/nba-pre-draft-measurements/.

            Just understand that I’m not going to touch on that because a sample size of 10 point guards or 20 wings is highly susceptible to noise.

            ” I’ll tell you what: there’s a reason why the “white” sample size is so small…”

            Because average-sized white athletes are pushed into other sports that are more white-dominated, whereas tall athletes (6’8″+) face a competitive disadvantage in almost all other sports, so basketball is the sport of choice regardless of race.

            I argued this in the article.

          • Layne Vashro

            By your argument, the sample of bigs should mirror the general US population. That means roughly 6 white big men for every black big man. However, your sample has nearly 4X as many back men as white men.

            I do not doubt that your theory addresses some of the differential representation in the NBA, but as a full explanation, it is not consistent with your data.

          • Will Garcia

            As a full explanation of the entire sample of bigs? DEFINITELY NOT!!! If it came across that way, I am very sorry for not communicating more clearly. Obviously, there are other factors at play here too.

            However, as an explanation for why the sample of white players is skewed toward big men and away from guards and wings, I believe that it may make up a large portion of the representative differences. This is what I was trying to convey.

          • Layne Vashro

            (disclaimer) I don’t mean to get too crazy on this issue, but this plays into something I have been working on with respect to which European countries produce basketball players. I doubt anyone else is paying attention, so I am going to geek out a bit.

            ____________
            I have an alternative explanation, which explains both the dearth of white perimeter players and the smaller, but still disparate, difference in representation among bigs.

            Only the 4.88th SD (~1 in 1,000,000) of white-male population in terms of “basketball ability” make it into the NBA, while the 4.28th SD (~2 in 100,000) of the black-male population makes it. This assumption is simply calculated from U.S. and NBA demographics.

            For an average american man, 5’10”, all of that advantage needs to come in non-height things like jumping, reaching, dribbling and shooting to be NBA quality. For a 6’10” man, 4 SDs above average, most of that specialness is accounted for by height. There is only a 1 in a billion chance your other skills are equally extreme. Most NBA bigs are only sorta special in non-tall basketball skills.

            This means that NBA big men are mostly defined by something that is distributed roughly the same in black and white men (in fact black men are about 1″ shorter which matters at extremes like this, but I don’t want to follow that tangent).

            If there is anything like vertical leaping, shooting, court-vision, or wingspan that is important to NBA success, but has a higher mean among the black population, we would expect the observed patterning of representation in the NBA by height. More white bigs, because height is the most heavily weighted factor in that set, but still an under representation because white men are less likely to be extreme in whatever other area is important.

          • Will Garcia

            So we’re arguing a pretty similar thing, except you’re saying that genetic/physical differences are doing most of the funneling, whereas I’m saying that it’s primarily sociocultural.

            That goes against what I’m seeing from Combine/Pre-Draft data, but could definitely be valid. Either way, it’s a sound explanation of how skills are distributed among sizes/positions.

  • http://www.twitter.com/The_Real_LLC Loren

    The Plumlee on the national team is Mason though.

    • Will Garcia

      Oh….darn.

      That’s…kind of embarrassing.

      Oh well, at least I didn’t mix him up with Marshall.

  • Pingback: The Black-White Thing: Racial Biases in NBA Scouting Comparisons – Hardwood Paroxysm | The Basketball Site

  • Old Bill Collins

    I am a white guy who has been following Pro Basketball since about 1950, that was before Russel and Chamberlain. The Globetrotters were arguably the best team even though they weren’t in the NBA. Our kids were in school and their team already could not play even with the teams from the black schools. The blacks had better genetic makeup for the game. They also lived in towns and cities where round ball was a year round game. That is similar to the Latino dominance in baseball. The year round play by young people sure develops superior skills. My family was supportive of the Civil Rights Movement and my one daughter choosing a black husband in the 60’s was not any problem for us.

    Having said that, I barely do notice skin color, but am very uncomfortable with much urban culture. The poor use of language and the vulgarity of the talk is not a good thing. The acceptance of tattoos to me is demeaning, almost like the branding of property. The noisiness and violence of the music makes me want to move away from the blacks. That is not a skin thing since I react the same to loud country music.

    The good thing is that roster spots are based on merit. Blacks have excelled and dominate the game. That is fair and I would have it no other way.