538 has a good piece on the implicit bias referees have when they get yelled at in NFL games:
a sideline bias in the NFL is real, and it’s spectacular. To prove it, we looked at the rates at which refs call the NFL’s most severe penalties, including defensive pass interference, aggressive infractions like personal fouls and unnecessary roughness, and offensive holding calls, based on where the offensive team ran its play.1…
For three common penalties, the direction of the play — that is, whether it’s run toward the offensive or defensive team’s sideline — makes a significant difference. In other words, refs make more defensive pass interference calls on the offensive team’s sideline but more offensive holding calls on the defensive team’s sideline. What’s more, these differences aren’t uniform across the field — the effect only shows up on plays run, roughly, between the 32-yard lines, the same space where coaches and players are allowed to stand during play.
Speaking as a referee, this makes intuitive sense. And it is a logical extension of the massive amount of research that shows crowd noise is a major factor in gaining home field advantage from refs. We’re human.
I would like to see a follow-up study for soccer and assistant referees. This would be a fairly clean study as the operational procedures produces a great data set for assistant referees. 95% of the time, both sets of benches are on the same side of the field. For the non-soccer folks, there are two assistant referees. AR-1 stands on the bench side with the right shoulder to the goal. A team bench is usually a few yards behind him and coaches have a technical area where they are allowed to wander freely. AR-2 is on the far side with no one behind them. Teams switch the direction of attack at half time.
So AR-1 has Team A in his ear for forty five minutes where A is attacking. And AR-1 also gets Team A in his ear for forty five minutes while they are defending. Team B does not have easy and constant access to AR-1 as they are always at least ten yards away from the halfline and at the professional level (where the data would be) there is a fourth official to act as a buffer.
My prediction is that Team A would over the course of the season have fewer offside violations called during its attack than Team B. I think the mechanism that will occur is that most assistant referees know that they are evaluated when the flag goes up on close calls. If they are not 100% certain that an offside violation has occurred, they are told to keep the flag down and not call the violation. If they miss an egregious offside, they will be graded down. But if they are not calling the occasional offside where the attacking player is off by half a shoe, their evaluation will not be impacted. None of this is conscious bias, it is human nature where a referee can firmly believe that they are only 95% sure instead of 100% and thus they keep their flag down.
I would love to see this data as I think the logic would hold true with a very clean data set.