The Univariate Fallacy is being used to forward the narrative that the police are "systemically racist" by presenting figures that show blacks are more frequently killed by police than one would predict by chance alone. 
This methodology is overly simplistic, since it fails to account for other variables, such as differences in crime rates, class, etc. In other words, univariate statistics are being applied to a complex, multivariate phenomenon, leading us to erroneous conclusions.
As @EPoe187 points out in this thread, males are shot by police at a MUCH higher rate than females, yet this is not evidence that the police are "systemically sexist" because we're all aware that other important variables are at play here.
This fact does not excuse police abuse and brutality, nor does it mean police reform isn't necessary. Nor does it mean, by any stretch of the imagination, that racism isn't a real problem in America. It is. But our beliefs and actions on any issue must be driven by data.
