Five years ago this month, the NBA released the Pedowitz Report, a 133-page tome authored by former federal prosecutor Larry Pedowitz that explored the extent to which real or perceived bias among officials permeates professional basketball. Like baseball’s Dowd Report in 1989, the NBA study was commissioned in response to a gambling scandal: NBA referee Tim Donaghy was found to be wagering on NBA games, including some that he officiated.
The Pedowitz Report suggested that “a substantial number of team representatives believe that referees make calls, on occasion, based on personal bias [and] the potential for referee bias remains a threat to the integrity of the game.” A year after the Pedowitz Report was released, Donaghy emerged from an 11-month stay in a federal penitentiary with a book alleging different types of bias—sometimes aimed at specific players, other times at team personnel—among select former colleagues. The most prominent example, as Donaghy explained in detail during a 60 Minutes appearance, involved certain referees having it in for Allen Iverson of the Philadelphia 76ers. For some fans, Donaghy’s book only served to strengthen a long-standing suspicion. Others ceased to trust the NBA officiating apparatus entirely—an attitude that’s still pervasive today.
The Iverson-related claims were quickly debunked, but a number of statistically testable allegations remained, some of which pertained to players, and others of which related to coaches and team owners. In three recently published academic articles, I studied several narrowly focused allegations of bias involving certain NBA referees against a specific player (Tim Duncan), a coach (Pat Riley), and an owner (Mark Cuban).
I found nothing.
In other words, I found no systematic evidence of a referee’s personal animus negatively impacting the performance of the team in question.
In statistical parlance, my non-findings are known as null results. Researchers often struggle to publish such findings, creating what is commonly known as the “file drawer problem”—that is, the problem in which the public only sees the studies with the most eye-catching findings. But the dissemination of null results is critically important in a number of realms, including tests of NBA referee impartiality. A finding of “no bias” among on-court officials is what one should find if the referees are properly neutral. In other words, perhaps NBA referees deserve more credit than they’ve been getting.
In the case of San Antonio Spurs star Tim Duncan, a distasteful on-court incident in April 2007 resulted in Duncan being ejected from the game in question and fined $25,000 for verbal abuse of referee Joey Crawford. For his role in the spate, Crawford was suspended for the remainder of the 2006-2007 season. The altercation gave rise to lingering fears of bias and retribution, and Duncan opined that Crawford had a “personal vendetta” against him. So I aimed to test whether any alleged vendetta actually manifested itself by detrimentally impacting the Spurs.
To do so, I looked at the 66 Spurs games Crawford officiated as a sub-set of the 1,000-plus games the Spurs played during 11 NBA seasons. What I found was that while the Spurs performed slightly worse than predicted during the 66 Crawford-officiated games, the results were not even close to being statistically significant. If Crawford was biased against Duncan and the Spurs team, it was not revealed on-court in any systematic way.
I used the same methodology and got similar results when I investigated allegations of referee bias made by Dallas Mavericks owner Mark Cuban and long-time Lakers, Knicks, and Heat coach/executive Pat Riley. Both men made a number of allegations; Riley’s comments pertained to specific referees, while Cuban’s remarks were aimed at NBA officiating in general. But in both cases, a systematic analysis of aggregate data over the course of multiple years revealed little. For example, despite Riley's pointed comments directed at now-retired NBA official Steve Javie, my analysis showed that over a nine-year period, his Miami Heat actually performed slightly better than predicted when Javie was one of the three referees on the court. Likewise, while my Mavericks-related study revealed the same playoff-only results others have found (that the team loses a curiously high number of playoff games when Danny Crawford referees), any negative bias directed against Cuban’s team disappeared when all games, both playoffs and regular-season, were analyzed in the aggregate.
I concede that my analytical model isn’t sensitive enough to pick up trace levels of bias at the individual game level. There is the insidious possibility that a referee who is consciously biased would reserve his or her ire for only the most critical moments during important games, thereby diluting any big-picture type of measure. I also recognize that the presence or absence of NBA referee bias plays a role among professional sports bettors, lending credence to such inquiries. In the wake of the 2007 NBA gambling scandal, the league implicitly acknowledged such a role when it moved to publish the identities of each game’s officials every morning on game day. Before that, the league kept referee information secret until only a couple of hours before the game, an act that served to create the specter of a market for inside information. Multiple experts acknowledge that referees can have a meaningful impact on the total points scored during a game, an outcome that can be wagered on in the form of over/under bets.
Certain claims of bias among basketball officials, though, can’t be empirically tested. They will always remain the topic of banter among suspicious sports fans. But for those allegations that can be researched, it’s important to disseminate all results—even those of the null variety. In addition to helping inform decision-making among various basketball constituents, such (non-)findings also serve to defend the integrity of NBA referees, a group that will be under close scrutiny as the 2013-2014 NBA season kicks off this week.