Is ref Danny Crawford biased against the Dallas Mavericks? And why isn’t the NBA smart enough to run basic statistical analyses?

The NBA assigned referee Danny Crawford to work tonight’s Dallas Mavericks/Portland Trailblazers game, touching off a series of galloping hissy fits across Greater East Texas/Shreveport metropolitan area. The reason for the unhappiness can be found in this paragraph. See if you can spot it.

The Mavs have a 2-16 record in playoff games officiated by Crawford, including 16 losses in the last 17 games. Dallas is 48-41 in the rest of their playoff games during the ownership tenure of Mark Cuban…

Wow. 11% with Crawford in the gym, 53% when he’s somewhere else. At a glance, that does look suspicious, doesn’t it? So I pinged my buddy Dr. Michael Pecaut of Pecaut Lab and Loma Linda University and asked if he’d run the stats for me. He did, and the results are about what I’d expected: these numbers are statistically significant at the .05 level. Which means that you’d only get this kind of variance by random chance one time in 20.* (Actually, Pecaut’s analysis says that the variance is actually higher than .05, so the odds are less than 5%, but that will work for purposes of this discussion.)

Okay. So what does this mean, in practical terms. I see three possibilities.

1. It could be random chance. There’s a one in 20 (or whatever) chance. Over the next 20 games Crawford calls involving the Mavericks, they could win 50-55% of the time, which would lend credence to the idea that we’re seeing an anomaly here. Or it could mean that this story becoming public caused him to reflect on his performance in Mavericks games and make changes to how he approaches them. Hard to say unless we have someone who can read minds.
2. It is theoretically possible that Crawford is calling the Mavericks straight-up and the NBA’s other 50-60 refs are badly biased in Dallas’ favor. If so, and we had a way of replacing fallible human refs with infallible refbots, we’d expect the Mavericks win percentage to be closer to 11% than 53%.
3. Or it’s possible that Crawford is biased against Dallas, either intentionally or unintentionally. The numbers don’t allow us to accuse the man of anything untoward, obviously, but they do allow us to assert that something is up and to suggest that we find out what it is.

The league has faith in Crawford:

“We have no concerns about the fairness of Danny Crawford’s officiating in Dallas games, whether in the playoffs or regular season,” NBA senior vice president of referee operations Ron Johnson said via email. “And there has never been any consideration given to not scheduling Crawford or any other referee to a particular team’s games.”

Of course, we’ve seen how the NBA handles criticism of its officiating in the past, so this is about what we’d expect them to say if a ref were caught on video fixing the game with the Gambinos at halftime.

What to Do, What to Do?

The NBA would do well to wake its dumb ass up. Those numbers are fact. They either mean something really bad or they appear to mean something bad in a way that defies easy explanation. Now they’re out in public. And the league’s fans already have plenty of reason to suspect the worst about its officiating. My guess is that if you ran similar tests on all of the league’s refs you’d find statistical indicators of potential bias in a number of places. It might not always be as stark as what we see in the Crawford/Mavericks case, and statistical variation might not always mean conscious bias, either.

If the league were bright, it would be running these kinds of analyses instead of waiting for ESPN to do it for them. And when they detect a potential issue – say, a ref’s results with a particular team vary at the .75 level, for instance, you pull the ref in for a talk. You don’t accuse, you just make the ref aware. If the refs are honest, as the league assumes, then they’re going to appreciate warning signs suggesting that their objectivity needs a tuneup. If a ref is working out a grudge, well, this kind of program is just the thing to scare him or her straight.

In the meantime, you might want to keep refs whose impartiality is statistically in question away from the teams they appear to have issues with, especially, you know, during the playoffs. Just saying.

UPDATE: As of this moment, Dallas is leading Portland by seven in the third period and Crawford just called three seconds on the Blazers.

(Notice how I did a whole post on NBA officiating and didn’t once mention Tim Donaghy?)

_____

* If you’re a stats wizard and want to subject this case (or any other involving NBA refs) to some analysis, let me know.