“Chicago police were 20 to 50 percent more likely to give sticker citations to vehicles parked in black neighborhoods compared to vehicles parked in non-black neighborhoods. The effect is robust to controls for annual employment levels and is not driven by … compliance."
Further, much of Fryer’s analysis relies on officers' self-reporting a suspect’s race. As @eluhper has convincingly shown for traffic stops in Texas, “misreporting was used effectively to evade detection of bias."
“Scrutiny may induce law enforcement officers to purposefully misreport their interactions with citizens to appear less biased.” Finds 60% of troopers misreported a minority as white when a searched failed to yield any evidence of a crime.