The one that found that refs (regardless of race) called things against players of other races more than against players of their own race.

There are also a lot of statistical…oddities in the playoffs. In 2007 & 2008, there were 25 series that went at least 5 games. When the home team was up 3-1, the visiting team shot 1.1% better. When the series was tied or the visitors were up 3-1, the home team shot 5.4% higher.
In Game 6, if the home team was behind in the series, they were called for 4.1 fewer fouls than the home team.
Game 7? No statistical oddities (foul differential of less than 1 between home and visitors). It suggests (but certainly doesn’t prove) that the NBA has attempted to stretch out playoff series whenever possible.

There’s also a Price/Remer/Stone paper that was published in March out of Oregon State (the three are from BYU, OSU, and Johns Hopkins) that finds statistical evidence of homecourt bias, refs attempting to keep games close, and playoff series extension, although they also say it doesn’t appear to be economically tied, and may be an unconscious bias on the part of the officials:
“Each type of favoritism results in approximately a 5-10% advantage in discretionary turnovers,and the first two types of bias (favoritism of home teams and teams losing during games) results in an approximate 2-5% advantage in shooting fouls.”